Wednesday, 2014-07-09

SukhdevSlickNik: There is one tricky step - you have to be very careful00:02
jeblairjhesketh: client.active_connections[0].sendAdminRequest(req)00:03
*** marcoemorais has joined #openstack-infra00:03
*** sabeen1 has quit IRC00:03
jheskethjeblair: thanks :-)00:03
*** zehicle_at_dell has quit IRC00:03
SukhdevSlickNik: Once you upload the plugin, and modify the config file (xml) file, you MUST NOT save config on jenkin's GUI00:03
clarkbzaro: see my comment on that change. I think you may want to address it in your change00:03
*** zehicle_at_dell has joined #openstack-infra00:04
SukhdevSlickNik: that will override the config file and you will loose the manually created config00:04
SukhdevSlickNik: That may be happening in your case00:04
*** dstufft|laptop has quit IRC00:04
clarkbzaro: let me know if you have questions about that, but zuul and other projects should have ltos of examples using python fixtures lib00:05
jeblairjhesketh: i think it would be useful to have bothe the client and worker id's and hosts (so 4 bits of info)00:05
SukhdevSlickNik: So, lets step back and verify in your system00:05
clarkbSukhdev: that sounds like a bug :)00:05
*** plars has quit IRC00:05
Sukhdevclarkb: yes, that makes this step very tricky00:06
SukhdevSlickNik: So, check the config file - open it and verify against your Jenkins setups00:06
SukhdevSlickNik: most likely your manual configuration for triggerOnEvents is gone00:07
SukhdevSlickNik: All you have to do is to retype that back in the config file - and restart Jenkins00:08
SukhdevSlickNik: you should be good to go00:08
*** anode has quit IRC00:08
SlickNikSukhdev: I just checked the config to make sure; and the commentPattern is still there.00:08
jheskethjeblair: okay, any particular order that makes sense to you?00:08
*** marcoemorais has quit IRC00:08
*** msabramo has quit IRC00:09
*** marcoemorais has joined #openstack-infra00:09
SukhdevSlickNik: Good, and the rest of the settings are also correct?00:09
*** marcoemorais has quit IRC00:09
*** marcoemorais has joined #openstack-infra00:09
*** dstufft|laptop has joined #openstack-infra00:10
*** marcoemorais has quit IRC00:10
SlickNikSukhdev: yes, looks like it.00:10
SukhdevSlickNik: Also be sure that all the xml tags match00:10
*** marcoemorais has joined #openstack-infra00:10
SlickNikSukhdev: Just want to check; this has been working correctly on the new version of gerrit (the one infra upgraded to a couple of months ago)?00:10
SukhdevSlickNik: or just cut and paste (as is) from my doc00:10
*** sweston has joined #openstack-infra00:11
*** msabramo has joined #openstack-infra00:11
*** plars has joined #openstack-infra00:11
jeblairjhesketh: nope00:12
SukhdevSlickNik: good question, I was not aware of this upgrade00:12
SlickNikSukhdev: I used the values from your doc — and verified that they're the ones that the plugin expects.00:12
clarkbSlickNik: it should, the api is the same00:12
clarkbzuul kept on working for example00:13
*** marcoemorais has quit IRC00:14
*** marcoemorais has joined #openstack-infra00:14
SlickNikclarkb: Sukhdev's patch is based off of is an older version of the gerrit-trigger-plugin, though (from sometime in February). I'm wondering if other fixes have gone in to the plugin to make it work for events since then, and they're not part of the patched plugin.00:15
SlickNikSukhdev: Can you quickly check if a comment-based recheck is working as expected for Arista CI?00:17
openstackgerritKhai Do proposed a change to openstack-infra/jenkins-job-builder: tests for JJB command processing
*** dims__ has quit IRC00:20
*** dstufft|laptop has quit IRC00:20
SlickNikSukhdev: I'm not sure if you're using a different regex, but a quick test "recheck no bug" on didn't trigger the Arista CI; so I have a feeling that it might be broken for you too.00:20
*** msabramo has quit IRC00:21
openstackgerritKhai Do proposed a change to openstack-infra/config: puppetize installation of gerrit third party plugins
*** wenlock has quit IRC00:24
SukhdevSlickNik: let me check00:25
Sukhdevclarkb: what was changed in the upgrade?00:27
clarkbSukhdev: we went from gerrit 2.4 to 2.800:27
clarkboh I know00:27
clarkbso gerrit includes more comment message data in the events now00:27
*** msabramo has joined #openstack-infra00:27
clarkbsomething like patchset X\n\nrecheck no bug00:27
clarkbthats probably it00:28
fungiyeah, we updated our zuul layout.yaml to accommodate that00:28
fungivote metadata now ends up in the start of the comment body00:28
Sukhdevclarkb fungi SlickNik: that is probably it then -00:28
*** msabramo has quit IRC00:32
*** rcarrillocruz has quit IRC00:35
SlickNikclarkb / sukhdev / fungi: I think that's part of it. I updated the comment regex for the plugin to be non-restrictive (.*), and it would still not take.00:35
SlickNikThere were some sections of the plugin that expect a verified=0 (or something like that) but I wasn't seeing any value for verified when I inspected the event json.00:36
openstackgerritKhai Do proposed a change to openstack-infra/config: Setup a private gerrit instance for security reviews
clarkbSlickNik: maybe you need multine regexes00:36
clarkb.* won't match \n by default00:36
clarkbin most re engines00:37
*** pballand has quit IRC00:37
*** sarob has quit IRC00:38
*** rcarrillocruz has joined #openstack-infra00:39
SlickNikclarkb: It's possible. I tested it here ( and it seemed to work. But I should probably dig into the code to see what sort of regex matching it's actually doing based on the input flags..00:40
*** tsg has joined #openstack-infra00:40
SlickNikclarkb: I'll look into that. Thanks for the suggestion!00:43
*** dims__ has joined #openstack-infra00:44
SlickNikfungi / clarkb / Sukhdev: Need to go afk for a bit, but will get to looking into this later tonight.00:47
*** pballand has joined #openstack-infra00:47
SlickNikThanks for the help!00:47
*** pcm_ has joined #openstack-infra00:47
*** pcm_ has quit IRC00:48
*** pcm_ has joined #openstack-infra00:48
*** regXboi has quit IRC00:49
*** leonchio_ has quit IRC00:49
*** salv-orlando has quit IRC00:50
*** dstufft|laptop has joined #openstack-infra00:54
*** dstufft|laptop has quit IRC00:54
*** shashankhegde has quit IRC00:55
*** penguinRaider has quit IRC00:55
*** msabramo has joined #openstack-infra00:57
*** penguinRaider has joined #openstack-infra00:57
*** marcoemorais has quit IRC00:59
*** amalagon has quit IRC01:01
*** julim has quit IRC01:03
*** amcrn has quit IRC01:06
*** bdpayne has joined #openstack-infra01:06
*** pballand has quit IRC01:09
*** mmaglana has quit IRC01:11
*** penguinRaider has quit IRC01:14
*** wenlock has joined #openstack-infra01:14
*** pcm_ has quit IRC01:16
*** penguinRaider has joined #openstack-infra01:16
*** Daisy has joined #openstack-infra01:18
*** penguinRaider has quit IRC01:19
*** penguinRaider has joined #openstack-infra01:19
*** obondarev has joined #openstack-infra01:19
*** bdpayne has quit IRC01:23
*** aysyd has quit IRC01:23
*** wenlock has quit IRC01:24
*** tsg has quit IRC01:28
*** dimtruck is now known as zz_dimtruck01:31
*** Ryan_Lane has quit IRC01:36
openstackgerritA change was merged to openstack-infra/devstack-gate: Stop doing tricky things with mirrors
*** msabramo has quit IRC01:37
*** rkukura has quit IRC01:43
*** nosnos has joined #openstack-infra01:44
*** zhiyan_ is now known as zhiyan01:47
*** arnaud has joined #openstack-infra01:47
*** moted has quit IRC01:48
*** amalagon has joined #openstack-infra01:51
*** Sukhdev has quit IRC01:51
*** msabramo has joined #openstack-infra01:53
*** homeless has quit IRC01:53
*** zhiyan is now known as zhiyan_01:56
*** zhiyan_ is now known as zhiyan01:56
*** bdpayne has joined #openstack-infra01:59
*** asettle has quit IRC01:59
*** markwash has quit IRC02:01
*** tsg has joined #openstack-infra02:01
openstackgerritMatthew Oliver proposed a change to openstack-infra/config: Update to instance variable notation in templates
*** sweston_ has joined #openstack-infra02:03
*** otter768 has joined #openstack-infra02:10
*** asettle has joined #openstack-infra02:12
*** jamielennox|away is now known as jamielennox02:12
*** jyuso has joined #openstack-infra02:16
*** pballand has joined #openstack-infra02:17
*** bdpayne has quit IRC02:18
*** bdpayne has joined #openstack-infra02:21
*** baohua has joined #openstack-infra02:22
*** bdpayne has quit IRC02:22
*** arnaud has quit IRC02:25
*** changbl has joined #openstack-infra02:30
openstackgerrittakehirokaneko proposed a change to openstack-infra/config: Add new stackforge project, rack
*** oomichi has joined #openstack-infra02:45
*** harlowja is now known as harlowja_away02:50
*** pballand has quit IRC02:51
*** signed8bit_ZZZzz has quit IRC02:52
*** torgomatic has left #openstack-infra02:52
*** MarkAtwood has joined #openstack-infra02:56
*** MarkAtwood has quit IRC02:56
*** signed8bit has joined #openstack-infra02:58
*** mmaglana has joined #openstack-infra02:58
*** sweston has quit IRC02:58
*** sweston_ has quit IRC02:59
*** arnaud has joined #openstack-infra03:01
*** shardy has quit IRC03:01
*** praneshp has quit IRC03:10
*** adalbas has quit IRC03:17
*** dstanek_zzz is now known as dstanek03:20
*** nosnos has quit IRC03:20
*** arnaud has quit IRC03:20
*** nati_ueno has quit IRC03:22
*** KarenZhu has joined #openstack-infra03:23
*** arnaud has joined #openstack-infra03:27
openstackgerritKhai Do proposed a change to openstack-infra/config: puppetize installation of gerrit third party plugins
tchaypoclarkb: no longer using pypi-mirror?03:40
*** pcm_ has joined #openstack-infra03:40
*** mestery has quit IRC03:40
tchaypoclarkb: have I missed a memo?03:40
clarkbtchaypo: you missed the memo03:41
tchaypois there a link to the memo?03:41
clarkbI think we mostly ninja'd it, but I posted about it in reply to jogo on a ml thread03:41
clarkbI think mordred will do a more formal RFA of it03:41
tchayponot that I object - is a good example of why it's a lot of effort to maintain03:42
uvirtbotLaunchpad bug 1294381 in openstack-ci "pypi-mirror does not pickup all necessary python packages" [Medium,Won't fix]03:42
tchaypoI think that we need to go ahead and remove the element and tweak tripleo docs so that we don't tell people to use it03:42
*** arnaud has quit IRC03:42
tchaypodo we have a suggested replacement?03:42
greghaynesIt wolrks fine for tripleo03:42
tchaypoI'm thinking bandersnatch..03:42
clarkbwell its ok to use it, we just won't be dealing with it03:42
clarkbbandersnatch is what we are using03:42
greghaynesIt actually works better for tripleo than bandersnatch, IMO03:43
tchaypogreghaynes: it's unsupported. Unless we plan to take on care and feeding I think we need to deprecate it.03:43
*** arnaud has joined #openstack-infra03:43
greghaynesIts unsupported?03:43
tchayposee clarkb's ninja move in
uvirtbotLaunchpad bug 1294381 in openstack-ci "pypi-mirror does not pickup all necessary python packages" [Medium,Won't fix]03:43
tchaypoI agree that it's a better fit though.03:44
clarkbtchaypo: in fairness I basically wrote the patch to fix that bug03:44
clarkband no one took me up on it03:44
clarkbso meh03:44
clarkbsnooze you lose or whatever03:44
tchaypoEspecially for new devs. Telling them "you need to download 30b of pypi as well as your OS archive" is...03:44
tchaypowell, a bit more effort than just using pypi-mirror03:44
*** KarenZhu has quit IRC03:45
*** lttrl has quit IRC03:45
greghaynesclarkb: any idea if there are other wontfix bugs for pypi-mirror?03:47
greghaynesor is there some lp incantation for finding that im unaware of...03:47
greghaynesoh, tags!03:47
clarkbI am not aware but there may be03:47
lifelessStevenK: ^03:47
clarkband yes tags is where to look03:47
StevenKtchaypo: 30 bits doesn't sound terrible :-P03:47
clarkbI just cleaned up the one I had been watching because :(03:48
lifelessbandersnatch is awesome03:48
lifelessI run both bandersnatch and pypi-mirror03:48
lifelesspypi-mirror solely for the wheel building03:48
lifelessclarkb: how do you address having wheels now ?03:49
clarkblifeless: currently we dont03:49
clarkbwe never used the wheels so it is not an issue for us03:49
*** otter768 has quit IRC03:49
StevenKlifeless: Do you log your bandersnatch mirror pulses?03:49
clarkbits possible mordred may end up wanting to do something like you are doing but for the time being its a non issue03:49
lifelessStevenK: probably03:50
StevenKlifeless: I see 50GiB is roughly what it takes to seed the mirror, I'm just wondering what the average daily churn is03:50
clarkbfwiw mordred does intend on doing the RFA thing iirc03:50
clarkband there will be more fanfare and all that03:50
tchaypoStevenK: you're terrible, kowalksiel03:50
clarkbit just happened it was bug day today so update the bugs03:50
clarkbStevenK: I thought mordred say closer to 90GiB03:51
*** otter768 has joined #openstack-infra03:51
mordredlifeless: does using pypi-mirror for wheel building work well?03:51
lifelessmordred: it makes image builds a hell-lot faster03:51
lifelessmordred: I've not had issues with it, since the pypi-mirror code fails-open03:52
lifelessStevenK: I have the logs. They don't have bw counts.03:52
greghaynescan also confirm - its a huge time win for me to have wheel mirror03:52
StevenKclarkb: I have a 12 hour window from midnight to midday that doesn't count to my quota03:52
* greghaynes wonders if this fix is really as easy as clarkb says it is ;)03:52
mordredlifeless: should we add an option to it to say "hey man, just build me some wheels"03:52
clarkbgreghaynes: it is03:53
lifelessmordred: you added that already03:53
mordredI did?03:53
clarkbStevenK: you have quotas? i am sorry03:53
lifelessmordred: I thought so, its a separate pass and creates a separate repo03:53
StevenKclarkb: Australia, what can you do03:53
clarkbmordred: yes pypi-mirror builds mirrors03:53
clarkber wheels03:53
mordredlifeless: its not a separate pass03:53
clarkbmordred: its the reason I pointed out that as being a thing to consider with the bandersnatch switch03:53
clarkbbut we didn't use them so we were all ok03:53
mordredI know it makes wheels03:53
lifelessmordred: well, so - theres a bunch of detail I just don't think about at the moment03:53
StevenKclarkb: 60GiB peak quota. Midnight to midday is 'off-peak' so is 'unlimited'03:54
mordredit also makes tarball mirror03:54
lifelessmordred: because its mirroring from my local bandersnatch03:54
StevenKclarkb: Sadly, 'unlimted' doesn't mean without limits.03:54
lifelessmordred: so I never see link failures etc03:54
* mordred will investigate03:54
lifelessmordred: I have it running in a container, and haven't thought about it in about 2 months03:54
clarkbright so I think a futuer thing would be to use bandersnatch mirror to build wheels similar to how pypi-mirror works03:54
tchaypoclarkb: thanks for updating the bug and being clear about what you're doing, it lets us react appropriately :)03:54
clarkbbut we don't need 75% of pypi-mirror to do that03:54
zaroclarkb: can you help me with this puppet error?
clarkbtchaypo: no problem :)03:55
lifelessI think what we're getting at is that perhaps we could trim pypi-mirror down and everyone will be happy03:55
*** dims__ has quit IRC03:55
*** asettle has quit IRC03:55
lifelessparticularly hard disk manufacturers03:55
mordredit's actually SUPER easy to do the wheel build03:56
mordredfrom a file03:56
jesusaurustrim pypi-mirror down to what?03:56
lifelessthe awkward thing03:56
lifelessis that we really want to build wheels for the transitive closure of deps in requirements03:56
mordredpip wheel -r requirements.text will do everything needed03:56
zaroclarkb: i don't see why it's it's looking for '/home/gerrit2/gerrit-plugins/javamelody-e00d5af.jar'. puppet command tells it to verify "/home/gerrit2/gerrit-plugins/javamelody.jar" exists.03:56
lifelessmordred: will it?03:56
lifelessmordred: transitively?03:56
lifelessjust checking cause, I'm so used to not having nice things03:57
StevenKI hope bandersnatch deals running as non-root03:57
lifelessStevenK: it does03:57
mordredwell, it may miss one ... but pypi-mirror will make the exact same mistakes03:57
mordredbut it will do transitive03:58
clarkbzaro: I think that happens because the exec to download the file doesn't download the file03:58
StevenKlifeless: Even when it tries to write out /etc/bandersnatch.conf ?03:58
mordredso trimming will be easy03:58
clarkbzaro: then the exec to install the file fails.03:58
clarkbzaro: I am not sure what the best way to fix that is03:58
lifelessStevenK: it's doing that cause it doesn't exist03:58
lifelessStevenK: its bonkers, but it only does it on missing03:58
mordredStevenK: bandersnatch upstream is very responsive if not03:58
annegentleha I didn't think reading TC meeting logs would cause me to ask for "Spoiler alert" on Germany/Brazil03:58
* mordred landed patch03:58
StevenKannegentle: Hahahaha03:59
lifelesswhich reminds me to nag dstufft again about pip releases.03:59
StevenKannegentle: You still haven't watched it?03:59
dstufftwhy do we want a pip 1.5.703:59
lifelessdstufft: ahha, its ALIVE03:59
*** arnaud has quit IRC03:59
lifelessdstufft: the fixes for local index urls IIRC04:00
lifelessstevebaker: ^ that was it, yes?04:00
dstufftoh I think that's in develop not 1.5.X04:00
lifelessdstufft: ok, so rephrasing. We wrote fixes, can we have them please :)04:00
tchaypoyeah, i think by the time it landed it would have had to be a 1.6.x04:00
tchaypofrom memory04:00
dstufftwe're not going to do a 1.5.7, but a 1.6 will be soonish04:00
*** otter768 has quit IRC04:01
dstufftnear future04:01
lifelessdstufft: days/weeks/?04:01
lifelessdstufft: the cause is, we have a bunch of tests we can't run until thats fixed, unless we setup a local http server etc04:01
clarkblifeless: well thankfully it is a solved problem :)04:02
zaroclarkb: how do you know the download failed?04:02
dstufftlifeless: start the release process in the next idk 2-3 weeks or so something like that04:02
clarkbzaro: because it is run in noop mode so the exec doesn't happen04:03
*** marun has joined #openstack-infra04:03
*** mmaglana has quit IRC04:03
lifelessclarkb: what is ?04:03
*** yamahata has joined #openstack-infra04:03
lifelessdstufft: ok, we'll put an HTTP server up04:03
clarkblifeless: case sensitivity and pip04:04
lifelessclarkb: so this isn't about case sensitivity :)04:05
*** arnaud has joined #openstack-infra04:05
*** matrohon has quit IRC04:05
lifelessclarkb: it was multiple bugs in the 404-equivalent handling of file:/// urls04:05
mordredit is for us04:05
lifelessit looks like case sensitivity issues04:06
clarkblifeless: right related to case sensitivity04:06
lifelessbut its really about 404s04:06
mordredfile:/// do case different04:06
clarkbpip with file:/// is case sensitive04:06
zaroclarkb: i still don't understanding why the error message references "javamelody-e00d5af.jar", should it not say 'could not find.. javamelody.jar'?04:06
clarkbpip with http:// is not case sensisitive04:06
lifelesscriky give me a second04:06
jesusauruszaro: clarkb: are you sure that "create" actually creates a File resource? the docs seem to indicate that it's more of an 'unless'04:06
lifelesscause I know why you are saying what you are saying04:06
lifelessbut you're actually observing a symptom of a larger issue, not what you think you are seeing04:06
lifelessand the issue is 404 handling04:07
lifelessand the reason that matters is if you're layering a partial mirror04:07
annegentleStevenK: watching it now, 40:30 or so :)04:07
mordredlifeless: awesome. the result is, we want your patch04:07
lifelesse.g. one containing just sdists of the things you want to test (e.g. just an sdist of os-apply-config)04:07
lifelessthen it will fail to fall over properly to the real index04:08
mordredannegentle: so you've seen the bits around 2504:08
annegentlemordred: oh yeah04:08
annegentlemordred: the sobbing fans are my favorites04:08
annegentleI'm notsogentle you know04:09
mordredannegentle: I have never seen anything like that04:09
clarkbits a semi final too04:09
clarkblike you expect this in the group games04:09
zarojesusaurus, clarkb : would it make sense to just remove the unless?04:09
jesusauruszaro: i would probably try changing the require from being a File to the Exec that creates the file04:10
clarkbjesusaurus: zaro nevermind line 75 is the problem04:10
clarkbjesusaurus: ya that04:10
mordredclarkb: not against Brazil in Brazil I don't04:12
clarkbmordred: sure. I think what I am trying to say is ghana drew germany04:13
clarkbso this was just insane04:13
*** nosnos has joined #openstack-infra04:13
tsgclarkb: continuing discussion on the failed py26 build (from earlier today)04:14
openstackgerritGregory Haynes proposed a change to openstack-infra/pypi-mirror: Force --no-use-mirror installaions
mordredclarkb: yah04:14
greghaynesclarkb: sorry for reopening your wontfix04:14
tsgclarkb: is there a way to reproduce the jenkins centos6 environment locally?  I am interested in looking at the requirements.txt that jenkins uses when it does "/home/jenkins/workspace/gate-swift-python26/requirements.txt, -r/home/jenkins/workspace/gate-swift-python26/test-requirements.txt"04:14
*** sweston has joined #openstack-infra04:15
clarkbgreghaynes: there are solutions to that problem :P04:15
clarkbtsg: thats the requirements.txt provided by your commit04:15
zarojesusaurus: is that with an Anchor?04:15
clarkbtsg: so nothing special04:15
tsgclarkb: gotcha .. how about global-requirements.txt?04:16
tsgis that prepopulated?04:16
tsgclarkb: I mean in the centos6 instance04:16
clarkbtsg: global-requirements shouldn't have anything to do with this test04:16
clarkbglobal-requirements only affect devstack installations04:16
*** sweston_ has joined #openstack-infra04:17
*** sweston__ has joined #openstack-infra04:17
*** sweston has quit IRC04:18
*** sweston has joined #openstack-infra04:18
jesusauruszaro: no, currently the install Exec requires the File specified by "creates" in the download Exec, but instead of requiring the File you should try requiring the download Exec04:18
clarkbtsg: if you spin up a centos6.5 vm and run `tox -epy26` that should get you pretty close04:19
jesusauruszaro: im guessing that the "creates" doesn't actually create a File resource in puppet dependency graph04:19
clarkbtsg: we install some additional packages on our slaves, you can install those with our puppet manifests if the above doesn't help04:19
clarkbjesusaurus: it doesn't04:19
tsgclarkb: ok :) is there a way to get the VM image for the slave?04:19
clarkbjesusaurus: you are right, its more like an unless04:19
clarkbtsg: there isn't04:19
clarkbtsg: bceause rax and hpcloud don't make that easy04:19
zarojesusaurus: having trouble finding an example of an exec in a require04:20
tsgclarkb: :) ok04:20
tsgclarkb: so stock centos6.5 + this recipe
clarkbtsg: the running your own there is for everything04:21
clarkbtsg: you can git clone cd config ; sudo ; sudo to get puppet installed04:22
clarkbthen you have to apply our thick slave template.04:22
clarkbbut I really don't think that is necessary04:22
clarkbI would start just with a boring centos6.5 image04:22
tsgclarkb: thanks, what does the thick slave template do04:22
clarkbtsg: it installs packages like libvirt and gettext04:23
jesusauruszaro: I commented with what you should replace that File with04:23
clarkbtsg: but since swift's dependency list is basically nil I don't think it makes much of a difference04:23
*** harlowja_at_home has joined #openstack-infra04:23
*** nosnos has quit IRC04:24
tsgclarkb: thanks much!  so essentially the "requirements.txt" jenkins refers to is only swift's own .. and not global-requirements.txt - if it weren't in global-requirements, it would still have worked ..04:27
tsgclarkb: I had the wrong idea all time (that getting into global-requirements is a must)04:27
openstackgerritKhai Do proposed a change to openstack-infra/config: puppetize installation of gerrit third party plugins
clarkbtsg: well it is04:27
clarkbtsg: but not for the unittests04:27
tsgclarkb: gotcha :)04:27
clarkbtsg: it is required in order to make changes to swift's requirements.txt merge04:27
anteayaSlickNik Sukhdev thanks for continuing to work on this04:28
clarkbtsg: is an example. I haven't gotten libvirt-python into global-requirements yet04:28
clarkbtsg: but I pushed a patch to nova to use it. The unittests pass but the integration tests do not04:28
tsgclarkb: ah .. the commit in question here actually add pyeclib to swift's requirements.txt04:28
tsgclarkb: just looked at your pending review .. see what you are saying04:29
clarkbtsg: right I would expect unittests to pass in that case. I wouldn't expect tempst to pass unless you had added pyeclib to global-requirements already04:29
clarkbso you must've done that step04:30
*** yfried__ has quit IRC04:30
tsgclarkb: yes, I had to go through the ordeal of adding py3 support to pyeclib :)04:30
tsgDoug wouldn't accept without it :)04:30
*** amitpp has joined #openstack-infra04:32
*** amitpp has quit IRC04:34
StevenKmordred, clarkb: Did you guys see ?04:35
*** praneshp has joined #openstack-infra04:35
openstackgerrittakehirokaneko proposed a change to openstack-infra/config: Add new stackforge project, rack
clarkbStevenK: mo04:35
clarkbjudging by the current logstash queue logstash is sucking04:37
clarkbso I think that rules out the problem being in elasticsearch client or binary protocol04:37
clarkbI think we should move forward with the local es daemon that logstash does http to. Then I can work on updating to latest logstash04:37
clarkbbecause the only thing better than the bug you know all of the new exciting bugs that come with new versions :)04:37
*** weshay has quit IRC04:38
*** amitpp has joined #openstack-infra04:38
*** garyk has quit IRC04:40
mordredachievement unlocked04:40
mordredpatch reviewed and approved on phone while at baseball game04:41
zarojesusaurus: yeah, finally jenkins+1.  thanks.04:43
jesusauruszaro: awesome04:43
clarkbmordred: nice. phones are amazing04:43
zaromordred: now i know why you want to review with email replies.04:44
*** elambert has joined #openstack-infra04:45
clarkbmordred: I got gertty running in screen on my irc box at one point04:45
clarkbbut didn't give it credentials04:45
clarkbnot sure it works so well on a phone04:45
zaroreminds me. gotta get it working on my mac before trip.04:46
mordredclarkb: we could totes do port of gertty to android ...04:46
clarkbmordred: you want to write more java? :P04:48
*** rkukura has joined #openstack-infra04:48
clarkbI am waiting for go to be on android04:48
mattoliveraumordred: get on that! +104:49
clarkbmattoliverau: careful, I am not sure what monty on java will look like. but it can't be pretty04:50
*** dims__ has joined #openstack-infra04:51
openstackgerritJoshua Harlow proposed a change to openstack-infra/config: Add templated uploading jobs for doc8 project
*** trinaths has joined #openstack-infra04:53
*** sweston has quit IRC04:54
*** sweston_ has quit IRC04:54
*** sweston__ has quit IRC04:54
*** dims__ has quit IRC04:56
*** dims__ has joined #openstack-infra04:57
*** sweston has joined #openstack-infra04:57
*** sweston_ has joined #openstack-infra04:58
*** sweston__ has joined #openstack-infra04:58
*** dims__ has quit IRC05:02
*** ildikov has quit IRC05:02
elambertclarkb: you've been helping Tushar (aka tsg) on a failure we are seeing in the swift-python26 environment. Quick question … do you happen to know if those centos6 images that run those jobs include the python-devel package?05:04
clarkbelambert: they do05:05
elambertok thanks05:05
*** timrc is now known as timrc-afk05:10
*** signed8bit has quit IRC05:11
*** praneshp_ has joined #openstack-infra05:11
*** signed8bit has joined #openstack-infra05:11
*** Ryan_Lane has joined #openstack-infra05:12
*** praneshp has quit IRC05:14
*** praneshp_ is now known as praneshp05:14
*** signed8bit has quit IRC05:16
*** harlowja_at_home has quit IRC05:17
*** timrc-afk is now known as timrc05:17
*** morganfainberg_Z is now known as morganfainberg05:18
*** chandankumar has joined #openstack-infra05:19
*** rcarrill` has joined #openstack-infra05:20
mordredclarkb: I wrote NDB/j back in the day to do low level access in java to mysql cluster05:22
*** rcarrillocruz has quit IRC05:23
clarkbmordred: its you05:23
clarkband Shrews!05:24
*** yfried__ has joined #openstack-infra05:36
*** nelsnelson has joined #openstack-infra05:41
*** chuckC has quit IRC05:41
*** ildikov has joined #openstack-infra05:42
morganfainbergmordred, NDB/j ? i might be a little frightened05:42
*** ominakov has joined #openstack-infra05:45
*** melwitt has quit IRC05:46
*** changbl has quit IRC05:51
*** garyk has joined #openstack-infra05:54
*** Ryan_Lane has quit IRC05:55
*** dims__ has joined #openstack-infra05:58
*** afazekas is now known as __afazekas05:58
*** Ryan_Lane has joined #openstack-infra06:03
*** dims__ has quit IRC06:03
mordredmorganfainberg: you should be very frightened06:04
*** reed has quit IRC06:06
*** xianghui has joined #openstack-infra06:07
*** dstanek is now known as dstanek_zzz06:08
*** mkoderer has joined #openstack-infra06:09
*** elambert has quit IRC06:11
openstackgerritMaru Newby proposed a change to openstack-infra/config: Fix neutron func job to use gate-specific tox env
*** morganfainberg is now known as morganfainberg_Z06:23
*** rdopieralski has joined #openstack-infra06:24
*** marun has quit IRC06:27
*** amalagon has quit IRC06:30
*** e0ne has joined #openstack-infra06:31
*** e0ne has quit IRC06:32
*** ominakov has quit IRC06:35
*** Longgeek has joined #openstack-infra06:36
*** e0ne has joined #openstack-infra06:36
*** rgerganov has quit IRC06:39
*** e0ne has quit IRC06:41
*** dstanek_zzz is now known as dstanek06:43
*** Ryan_Lane has quit IRC06:44
openstackgerritRadomir Dopieralski proposed a change to openstack/requirements: Add xstatic-packaged JavaScript libraries for Horizon
*** e0ne has joined #openstack-infra06:46
*** doude has joined #openstack-infra06:46
*** amalagon has joined #openstack-infra06:47
*** e0ne has quit IRC06:48
openstackgerritafazekas proposed a change to openstack-infra/devstack-gate: Neutron default tenant network type vxlan
*** mkerrin has joined #openstack-infra06:49
*** e0ne has joined #openstack-infra06:52
*** arnaud has quit IRC06:53
*** dstanek is now known as dstanek_zzz06:53
*** jcoufal has joined #openstack-infra06:55
*** amcrn has joined #openstack-infra06:56
*** dims__ has joined #openstack-infra06:58
*** flaper87|afk is now known as flaper8706:59
*** jistr has joined #openstack-infra07:01
*** dims__ has quit IRC07:03
*** xianghui has quit IRC07:04
*** luqas has joined #openstack-infra07:05
*** luqas has quit IRC07:07
*** skolekonov has joined #openstack-infra07:08
*** luqas has joined #openstack-infra07:10
*** mpavone has joined #openstack-infra07:15
*** msabramo has quit IRC07:16
*** e0ne has quit IRC07:18
*** jlibosva has joined #openstack-infra07:20
*** rgerganov has joined #openstack-infra07:23
*** afazekas has joined #openstack-infra07:25
*** tkelsey has joined #openstack-infra07:26
*** nibalizer has quit IRC07:27
*** jgallard has joined #openstack-infra07:28
*** mrda is now known as mrda-away07:28
*** e0ne has joined #openstack-infra07:33
*** pcm_ has quit IRC07:33
*** mpavone1 has joined #openstack-infra07:34
*** mpavone has quit IRC07:34
*** freyes has joined #openstack-infra07:36
*** jgallard has quit IRC07:36
*** jgallard has joined #openstack-infra07:36
*** hashar has joined #openstack-infra07:36
*** ominakov has joined #openstack-infra07:40
*** e0ne has quit IRC07:41
*** che-arne has joined #openstack-infra07:46
*** shardy has joined #openstack-infra07:47
*** yaguang has joined #openstack-infra07:53
*** masayukig has joined #openstack-infra07:54
*** ildikov has quit IRC07:55
*** ildikov has joined #openstack-infra07:56
*** markmc has joined #openstack-infra07:57
*** jpich has joined #openstack-infra07:58
*** dims__ has joined #openstack-infra07:59
markmcSergeyLukjanov, any idea why zuul isn't picking up ?08:04
*** pcrews has quit IRC08:04
*** dims__ has quit IRC08:04
*** xianghui has joined #openstack-infra08:06
*** amcrn_ has joined #openstack-infra08:07
*** amcrn has quit IRC08:08
*** amcrn_ is now known as amcrn08:08
*** Alexei_9871 has quit IRC08:09
*** derekh_ has joined #openstack-infra08:11
*** Guest58876 has joined #openstack-infra08:14
*** andreykurilin_ has joined #openstack-infra08:14
*** zz_johnthetubagu is now known as johnthetubaguy08:26
*** dims__ has joined #openstack-infra08:28
*** pblaho has joined #openstack-infra08:32
hasharcould it be missing another code-review +2 vote ?08:32
hasharah no08:33
*** penguinRaider has quit IRC08:34
*** dims__ has quit IRC08:34
*** persia has quit IRC08:37
*** persia has joined #openstack-infra08:37
*** persia has quit IRC08:37
*** persia has joined #openstack-infra08:37
*** jedimike has joined #openstack-infra08:39
*** hashar has quit IRC08:40
*** hashar has joined #openstack-infra08:40
rdopieralskiis Thomas Goirand here?08:47
Mithrandirrdopieralski: zigo08:49
zigoyes, I am.08:49
zigordopieralski: ^08:50
*** ihrachyshka has joined #openstack-infra08:50
SergeyLukjanovmarkmc, it's strange but it works ok after the "recheck no bug"08:57
*** jedimike has quit IRC08:59
rdopieralskizigo: would you have a moment to talk about ?08:59
rdopieralskizigo: I'm at a loss about what you are expecting from me there.08:59
bookwarhi, everyone. Where can i find merge requirements for stackforge subproject? is it configurable?08:59
*** jedimike has joined #openstack-infra08:59
*** tusharsg has joined #openstack-infra09:00
zigordopieralski: You're putting a way too high versions, which aren't even available in Debian.09:00
*** mpavone1 has quit IRC09:01
zigordopieralski: For example, jquery is version 1.7.2+dfsg-3 in Debian, so we shouldn't require something higher, unless there's a reason for it (like a new feature we need).09:01
zigoSame for all javascript libs.09:01
*** tsg has quit IRC09:02
rdopieralskizigo: I'm not putting those versions there, those are the versions that Horizon is already using, since Icehouse09:02
zigoThis is completely orthogonal to the version of xstatic in pypi, because Debian will change the version numbers to match whatever the version we have for the jscript lib in Debian.09:02
zigordopieralski: But are you *sure* we need something that high?09:02
zigoMaybe Horizon doesn't even use the new features in these versions...09:02
zigoI got to go, sorry.09:03
rdopieralskizigo: perhaps, but re-writing code to use older versions than the previous release would be quite some work09:03
openstackgerritA change was merged to openstack-infra/devstack-gate: No more thinging in
zigordopieralski: My bet is that it already works in many cases. That's at least truth for jquery.09:04
rdopieralskizigo: no, jquery 1.10 is not compatible with older versions, a lot has been changed09:05
zigordopieralski: Though, horizon still work !!!09:05
rdopieralskizigo: but that's beside the point, Horizon already shipped with those versions on last release09:05
rdopieralskizigo: I'm not proposing to upgrade the versions of those libraries, I'm merely cleaning up by removing them from Horizon's repository into separate packages09:06
rdopieralskizigo: if you think that Horizon should use older libraries, that's a completely separate issue and we would probably need a blueprint for that, and I doubt it would get done in this release cycle09:07
*** rcarrillocruz has joined #openstack-infra09:08
rdopieralskizigo: anyways, please catch me when you have some time to discuss this, maybe you I should start a thread on the mailing list?09:09
*** jp_at_hp has joined #openstack-infra09:10
*** e0ne has joined #openstack-infra09:10
*** rcarrill` has quit IRC09:10
*** andreykurilin_ has quit IRC09:12
markmcSergeyLukjanov, thanks09:13
*** mpavone has joined #openstack-infra09:15
*** _shaps_ has joined #openstack-infra09:16
*** pelix has joined #openstack-infra09:19
*** e0ne has quit IRC09:20
*** jyuso has quit IRC09:22
*** e0ne has joined #openstack-infra09:23
*** Longgeek has quit IRC09:27
*** Daisy has quit IRC09:29
*** dims__ has joined #openstack-infra09:30
*** dims__ has quit IRC09:35
*** sweston__ has quit IRC09:36
*** sweston has quit IRC09:36
*** sweston_ has quit IRC09:36
*** lcheng has joined #openstack-infra09:38
*** hashar has quit IRC09:43
*** yamahata has quit IRC09:48
*** dizquierdo has joined #openstack-infra09:50
*** praneshp has quit IRC09:56
*** ildikov has quit IRC09:57
*** Longgeek has joined #openstack-infra09:58
*** Guest58876 has quit IRC09:58
sdaguemarkmc: we've had a few of those recently, I wonder if zuul is losing events10:03
sdagueSergeyLukjanov: any chance you can take a look at the logs?10:03
*** _nadya_ has joined #openstack-infra10:05
*** Longgeek has quit IRC10:05
*** jistr has quit IRC10:06
*** _nadya_ has quit IRC10:06
*** jgallard has quit IRC10:08
*** tusharsg has quit IRC10:11
*** ildikov has joined #openstack-infra10:15
*** jerryz has quit IRC10:21
*** MIDENN_ has quit IRC10:23
openstackgerritNikhil Manchanda proposed a change to openstack-infra/config: Fixed gate-trove-functional-dsvm jobs
*** jistr has joined #openstack-infra10:27
*** penguinRaider has joined #openstack-infra10:30
*** dims__ has joined #openstack-infra10:31
*** ociuhandu has joined #openstack-infra10:35
*** alexpilotti has joined #openstack-infra10:35
*** e0ne has quit IRC10:35
*** dims__ has quit IRC10:36
SergeyLukjanovsdague, I've been from phone, now find my laptop and looking on logs10:36
*** e0ne has joined #openstack-infra10:37
*** Shrews has quit IRC10:43
*** habib has joined #openstack-infra10:43
*** Shrews has joined #openstack-infra10:47
*** yamahata has joined #openstack-infra10:47
*** rlandy has joined #openstack-infra10:49
*** xianghui has quit IRC10:49
SergeyLukjanovsdague, heh, I don't see any issues in ligs10:55
*** weshay has joined #openstack-infra10:57
SergeyLukjanovsdague, yeah10:57
*** trinaths has quit IRC10:58
SergeyLukjanovsdague, the only related errors in logs are10:58
SergeyLukjanov7905:2014-07-09 08:56:36,665 INFO zuul.Scheduler: Adding openstack/oslo.config, <Change 0x7fdbab60b110 105539,1> to <Pipeline silent>10:58
SergeyLukjanov7906-2014-07-09 08:56:36,666 ERROR zuul.IndependentPipelineManager: Unable to find change queue for project openstack/oslo.config10:58
SergeyLukjanov7907:2014-07-09 08:56:36,668 INFO zuul.Scheduler: Adding openstack/oslo.config, <Change 0x7fdbab60b110 105539,1> to <Pipeline check-tripleo>10:58
SergeyLukjanov7908-2014-07-09 08:56:36,669 ERROR zuul.IndependentPipelineManager: Unable to find change queue for project openstack/oslo.config10:58
SergeyLukjanovbut I don't think that it could break something10:59
*** ZZelle_ has joined #openstack-infra11:00
sdagueis there anything about Pipeline check?11:00
*** freyes has quit IRC11:00
openstackgerritPetr Blaho proposed a change to openstack-infra/config: Adds pip freeze output to pep8 job
*** xianghui has joined #openstack-infra11:03
*** freyes has joined #openstack-infra11:06
*** dims__ has joined #openstack-infra11:10
*** jcoufal-mobile has joined #openstack-infra11:13
*** rcarrill` has joined #openstack-infra11:21
*** dstanek_zzz is now known as dstanek11:22
*** rcarrillocruz has quit IRC11:23
*** lcheng has quit IRC11:24
*** alexpilotti has quit IRC11:26
*** yaguang has quit IRC11:27
*** alexpilotti has joined #openstack-infra11:30
*** lcheng has joined #openstack-infra11:31
*** otter768 has joined #openstack-infra11:32
*** rcarrillocruz has joined #openstack-infra11:33
*** tkelsey has quit IRC11:34
*** rcarrill` has quit IRC11:34
*** alexpilotti has quit IRC11:35
*** Longgeek has joined #openstack-infra11:38
*** denis_makogon has joined #openstack-infra11:42
*** otter768 has quit IRC11:42
*** pcm_ has joined #openstack-infra11:43
*** lcheng has quit IRC11:44
openstackgerritSergey Lukjanov proposed a change to openstack-infra/config: Limit sahara's stable acls to icehouse
openstackgerritSergey Lukjanov proposed a change to openstack-infra/config: Remove sahara's tag acls
openstackgerritSergey Lukjanov proposed a change to openstack-infra/config: Add specific python-saharaclient acls
SergeyLukjanovttx, ^^ cleanup of sahara's acls11:47
*** jcoufal-mobile has quit IRC11:49
*** NikitaKo_ has joined #openstack-infra11:49
*** dprince has joined #openstack-infra11:53
*** jistr is now known as jistr|english11:56
*** ominakov has quit IRC11:57
*** ominakov has joined #openstack-infra11:58
*** pcm_ has quit IRC11:59
*** pcm_ has joined #openstack-infra12:00
*** lcheng has joined #openstack-infra12:01
kashyapInterestingly on #openstack-nova is not logged? --
openstackgerritDavanum Srinivas (dims) proposed a change to openstack-dev/pbr: Switch from to
AJaegerkashyap, that's a team decision - not every channel is logged, #openstack-doc is not logged either12:05
openstackgerritSergey Skripnick proposed a change to openstack-infra/config: Fix file name and path of generated html results
*** dstanek is now known as dstanek_zzz12:11
*** jistr|english is now known as jistr12:11
*** lcheng has quit IRC12:11
*** doug-fish has joined #openstack-infra12:15
kashyapAJaeger, I see. Yeah, I just noticed as I wanted to link to a conversation, and found it wasn't logged. No worries.12:18
*** dstanek_zzz is now known as dstanek12:18
*** rcarrill` has joined #openstack-infra12:18
*** rcarrillocruz has quit IRC12:18
*** yaguang has joined #openstack-infra12:23
AJaegerkashyap, understood, thanks12:23
*** aysyd has joined #openstack-infra12:23
*** lcheng has joined #openstack-infra12:24
*** rcarrill` has quit IRC12:24
*** ArxCruz has joined #openstack-infra12:26
*** e0ne_ has joined #openstack-infra12:28
*** e0ne has quit IRC12:31
*** dims__ has quit IRC12:32
*** dims__ has joined #openstack-infra12:32
*** mwagner_lap has joined #openstack-infra12:36
*** hashar has joined #openstack-infra12:36
*** juzuluag has joined #openstack-infra12:36
*** adalbas has joined #openstack-infra12:38
*** prad has quit IRC12:40
*** lttrl has joined #openstack-infra12:40
*** mkoderer has quit IRC12:42
*** e0ne_ has quit IRC12:43
*** _nadya_ has joined #openstack-infra12:45
*** Daisy has joined #openstack-infra12:46
Shrewsclarkb: and now you know one more reason why i despise java12:46
*** e0ne has joined #openstack-infra12:48
*** pcm_ has quit IRC12:48
*** julim has joined #openstack-infra12:48
*** dkranz has joined #openstack-infra12:49
*** chuckC has joined #openstack-infra12:49
*** radez_g0n3 is now known as radez12:51
*** _nadya_ has quit IRC12:54
*** _nadya_ has joined #openstack-infra12:55
*** baohua has quit IRC12:56
*** mbacchi has joined #openstack-infra12:57
*** yfried__ has quit IRC12:58
*** _nadya_ has quit IRC12:59
*** sballe has joined #openstack-infra13:01
*** dangers_away is now known as dangers13:04
*** asselin has quit IRC13:07
*** asselin has joined #openstack-infra13:07
*** yaguang has quit IRC13:08
*** fanhe has joined #openstack-infra13:10
*** oomichi has quit IRC13:10
*** changbl has joined #openstack-infra13:10
*** zul has quit IRC13:11
*** signed8bit has joined #openstack-infra13:11
*** eharney has quit IRC13:14
*** ildikov has quit IRC13:14
*** habib has quit IRC13:15
*** amotoki has quit IRC13:16
*** mriedem has joined #openstack-infra13:18
*** Guest58876 has joined #openstack-infra13:18
*** zul has joined #openstack-infra13:18
fungimarkmc: sdague: it's!/story/14713:19
sdaguefungi: that says New Task13:19
fungimarkmc: sdague: no idea if my patch is "the right way" to fix that13:19
sdagueNo Title13:19
fungiit pulls it up for me... maybe you have to be logged into storyboard first?13:20
*** signed8bit has quit IRC13:20
fungiactually no, i wasn't even logged in at the time i checked that13:20
sdaguefungi: yeh, you have to be logged in13:20
sdaguethat's *terrible*13:20
fungimaybe i had a cookie or cache13:20
*** signed8bit has joined #openstack-infra13:20
*** pcm_ has joined #openstack-infra13:21
*** mestery has joined #openstack-infra13:22
sdagueyeh, it works for me now, no idea13:22
fungisdague: markmc: anyway, to cut the story short i've observed that when a parent change fails to merge and a change behind it gets dequeued, zuul can fail to report in some circumstances leaving no vote. then later when the parent is reapproved there's no +1 from zuul on the child change so it's discounted from automatically entering the gate13:22
sdaguewhen I went there it was completely blank though13:22
*** pdmars has joined #openstack-infra13:23
fungisdague: markmc: this is in contrast to zuul's previous behavior where it automatically reenqueued child changes into the gate when the parent was reapproved regardless of what its check results had been, so overall the behavior is an improvement. as a result i'm just attempting to address the reasons for zuul failing to report/vote13:24
*** tkelsey has joined #openstack-infra13:24
sdagueok, so this is mostly because zuul doesn't have history, right?13:24
*** Longgeek has quit IRC13:25
sdagueit needs to also know the last vote, not the current state?13:25
*** _nadya_ has joined #openstack-infra13:25
fungiyeah, though i think the behavior change is going to be more obvious once the reporting failure is fixed13:26
*** yaguang has joined #openstack-infra13:26
funginamely, that when zuul leaves a -2 on a change because it failed to merge due to the failure of a required parent change, it will not be reenqueued until it's manually rechecked13:26
fungior separately reapproved or whatever13:26
*** garyk has quit IRC13:27
*** masayukig has quit IRC13:27
*** wenlock has joined #openstack-infra13:28
*** jgrimm has joined #openstack-infra13:28
*** dustins has joined #openstack-infra13:31
*** mfer has joined #openstack-infra13:32
*** homeless has joined #openstack-infra13:33
*** msabramo has joined #openstack-infra13:33
*** masayukig has joined #openstack-infra13:34
*** Longgeek has joined #openstack-infra13:35
dhellmannfungi: can you add me to the oslo-db-ptl group in gerrit so I can let viktors do a release today, please?13:38
*** viktors has joined #openstack-infra13:38
*** saper_ has joined #openstack-infra13:39
*** dims_ has joined #openstack-infra13:39
fungidhellmann: done13:40
dhellmannfungi: thank you, sir!13:40
*** pcrews has joined #openstack-infra13:40
*** rickerc_ has joined #openstack-infra13:40
*** dstanek is now known as dstanek_zzz13:41
*** eglynn-regus has joined #openstack-infra13:41
*** yfried__ has joined #openstack-infra13:41
dhellmannfungi: IIRC, there were other group changes in that ACL patch I worked on with anteaya. Is someone auditing those, or should I make a sweep and report the ones I need to be added to? I'm happy to do it, but don't want to duplicate effort.13:41
*** davidlenwell_ has joined #openstack-infra13:41
*** duran has joined #openstack-infra13:41
*** yfried__ has quit IRC13:41
fungidhellmann: unless anteaya was going to make a list of what you needed to be added to, nobody's working on it that i know of13:42
dhellmannthere was an etherpad...13:42
*** yfried__ has joined #openstack-infra13:42
* dhellmann looks in bookmarks13:42
*** lifeless1 has joined #openstack-infra13:43
*** GheRiver1 has joined #openstack-infra13:43
*** jistr has quit IRC13:43
*** jamespd_ has joined #openstack-infra13:44
*** gilliard_ has joined #openstack-infra13:44
*** SpamapS_ has joined #openstack-infra13:44
*** cnesa has joined #openstack-infra13:44
*** dkranz has quit IRC13:45
openstackgerritJeremy Stanley proposed a change to openstack-infra/config: Delete select-mirror and all the args
dhellmannfungi: let me see if I can make that list shorter13:46
*** zz_dimtruck is now known as dimtruck13:46
fungidhellmann: yeah, and it looks like some changed names or something13:46
*** rkukura has quit IRC13:46
*** che-arne has quit IRC13:46
*** dansmith_ has joined #openstack-infra13:47
*** krotscheck_ has joined #openstack-infra13:47
dhellmannfungi: yeah, I'll make a list of the stuff I can't fix myself and ping you when I'm done with that13:47
*** mordred_ has joined #openstack-infra13:47
*** che-arne has joined #openstack-infra13:47
*** tsg has joined #openstack-infra13:47
*** tsg has quit IRC13:48
*** homeless has quit IRC13:48
*** tkelsey has quit IRC13:48
*** dims__ has quit IRC13:48
*** NikitaKo_ has quit IRC13:48
*** rlandy has quit IRC13:48
*** gema has quit IRC13:48
*** gilliard has quit IRC13:48
*** rickerc has quit IRC13:48
*** krtaylor has quit IRC13:48
*** nijaba has quit IRC13:48
*** dkehn has quit IRC13:48
*** greghaynes has quit IRC13:48
*** Mithrandir has quit IRC13:48
*** mordred has quit IRC13:48
*** uvirtbot has quit IRC13:48
*** jamespd has quit IRC13:48
*** vladan has quit IRC13:48
*** dansmith has quit IRC13:48
*** NobodyCam has quit IRC13:48
*** loquacities has quit IRC13:48
*** mattymo has quit IRC13:48
*** mordred_ is now known as mordred13:48
*** GheRivero has quit IRC13:48
*** jogo has quit IRC13:48
*** saper has quit IRC13:48
*** skraynev has quit IRC13:48
*** davidlenwell has quit IRC13:48
*** SpamapS has quit IRC13:48
*** lifeless has quit IRC13:48
*** krotscheck has quit IRC13:48
*** anteaya has quit IRC13:48
*** sdague has quit IRC13:48
*** dragondm has quit IRC13:48
*** eglynn-offic-afk has quit IRC13:48
*** dansmith_ is now known as dansmith13:48
*** krotscheck_ is now known as krotscheck13:48
*** rickerc_ is now known as rickerc13:48
*** jogo has joined #openstack-infra13:49
*** tsg has joined #openstack-infra13:49
*** homeless has joined #openstack-infra13:49
*** gema has joined #openstack-infra13:51
*** cnesa2 has joined #openstack-infra13:51
*** cnesa has quit IRC13:51
*** andreaf has joined #openstack-infra13:51
*** yamahata has quit IRC13:52
*** yamahata has joined #openstack-infra13:53
*** sdague has joined #openstack-infra13:53
*** rlandy has joined #openstack-infra13:54
*** anteaya has joined #openstack-infra13:54
*** skraynev has joined #openstack-infra13:54
*** NikitaKo_ has joined #openstack-infra13:55
fungimordred: if you agree with my comment on i'm happy to work on carving that up into separate phased changes13:56
*** nijaba has joined #openstack-infra13:56
*** nijaba has joined #openstack-infra13:56
*** dstanek_zzz is now known as dstanek13:56
*** tkelsey has joined #openstack-infra13:56
*** uvirtbot has joined #openstack-infra13:57
dhellmannfungi: ok, there's a list at the bottom of under "New and empty groups"13:57
openstackgerritValeriy Ponomaryov proposed a change to openstack-infra/config: Set py33 job for python-manilaclient as votable
*** Mithrandir has joined #openstack-infra13:57
*** baohua has joined #openstack-infra13:58
*** NobodyCam has joined #openstack-infra13:58
*** dkehn has joined #openstack-infra13:58
mordredfungi: yes. I agree13:58
fungidhellmann: when you say "could also remove/rename..." you mean that the group name changed in the acl, so the old group name is no longer in use?13:59
mordredfungi: and godo call13:59
*** prad has joined #openstack-infra13:59
dhellmannfungi: yes, that's right13:59
fungimordred: also check out the updates i made to the change if you haven't already and make sure they weren't things you left out on purpose13:59
*** dragondm has joined #openstack-infra13:59
mordredfungi: they were not14:00
fungiokay, great14:00
*** jistr has joined #openstack-infra14:00
fungidhellmann: we could have renamed the groups before applying the acl change, but at this point its easier for me to just recreate the membership in the new ones and empty/hide the old ones14:01
dhellmannfungi: ok, that works, I wasn't sure which was easier.14:01
*** cnesa2 has quit IRC14:01
*** loquacities has joined #openstack-infra14:02
*** mordred has quit IRC14:02
*** mordred has joined #openstack-infra14:02
*** vladan has joined #openstack-infra14:03
*** Daisy has quit IRC14:03
*** eharney has joined #openstack-infra14:03
*** jistr has quit IRC14:03
*** YorikSar has quit IRC14:04
*** jistr has joined #openstack-infra14:04
*** mattymo has joined #openstack-infra14:04
*** jcoufal has quit IRC14:06
*** pballand has joined #openstack-infra14:08
*** krtaylor has joined #openstack-infra14:09
*** NikitaKo_ has quit IRC14:09
*** jcoufal has joined #openstack-infra14:10
*** greghaynes has joined #openstack-infra14:11
*** garyk has joined #openstack-infra14:11
*** obondarev has quit IRC14:11
fungidhellmann: you're a direct member of each of those groups, oslo-ptl has been added as an included group in each, and the existing membership of the groups those were renames from have been duplicated in them as well (with the exception of any which would have been including themselves)14:11
fungifeel free to trim those as you see fit14:11
dhellmannfungi: thanks, that's great!14:12
*** amitpp has quit IRC14:12
fungimordred: we're seeing an increased incidence of these in the gate now that i finally got your d-g change merged...
*** gargola has quit IRC14:13
fungimordred: i'm going to look into possible resource constraints/tuning needed on the current mirror, but we may want to start thinking about adding local mirrors soonish14:14
*** alexpilotti has joined #openstack-infra14:14
fungimordred: i also haven't ordered the x.509 cert for pypi.o.o yet, but i think we should hold off switching it to https anyway until we're sure we're not going to push it over the edge performance-wise14:15
*** ihrachyshka has quit IRC14:16
*** mdenny has joined #openstack-infra14:16
mordredfungi: nod14:17
fungithough the socket timeouts may be entirely coincidental, since we were serving most of this off pypi.o.o already albeit from a reduced scope mirror14:17
mordredfungi: but we were only serving to half of the projects14:17
*** ihrachyshka has joined #openstack-infra14:17
fungiwe're basically adding some load from stackforge and infra projects, yeah14:18
*** dpyzhov has joined #openstack-infra14:18
mordredyah. but that's over half of the projects in the system14:18
*** dpyzhov has left #openstack-infra14:18
sdaguefungi: you have logs on pypi.o.o to figure out if you are actually overrunning the connections?14:18
fungithough if memory serves, they first checked the mirror for packages provided there and then checked, so we may have been serving a fair amount of their reqs from the mirror too14:18
fungisdague: i expect so. i'll know shortly14:19
fungiany guesses why this is leaking into the top-level apache error.log? [error] No handlers could be found for logger "keystoneclient.httpclient"14:22
mordredfungi: on static?14:22
*** msabramo has quit IRC14:23
JayFThis has 2x+1 and 1x+2 if someone could look at it and +2/+A:
fungimordred: yeah14:23
mordredJayF: looking14:23
hasharhello! I had a question with  recheck bug 12345 , does Zuul has supports to match that 12345 and pass it as a parameter to a job ?14:24
uvirtbotLaunchpad bug 12345 in isdnutils "isdn does not work, fritz avm (pnp?)" [Medium,Fix released]
fungioh, i bet the keystoneclient.httpclient bit is courtesy of the object storage backend stuff14:25
*** viktors has left #openstack-infra14:26
*** jaypipes has joined #openstack-infra14:27
*** dprince has quit IRC14:27
*** pblaho has quit IRC14:27
*** lttrl has quit IRC14:28
*** pdmars has quit IRC14:28
*** pdmars has joined #openstack-infra14:29
*** zehicle_at_dell has quit IRC14:29
*** tsg has quit IRC14:33
openstackgerritSean Dague proposed a change to openstack-infra/elastic-recheck: add glance deadlock query
*** msabramo has joined #openstack-infra14:34
*** pballand has quit IRC14:35
*** pcm_ has quit IRC14:35
*** pcm_ has joined #openstack-infra14:35
*** mkoderer has joined #openstack-infra14:37
*** ominakov has quit IRC14:37
*** gargola has joined #openstack-infra14:38
openstackgerritA change was merged to openstack-infra/elastic-recheck: add glance deadlock query
mordredfungi: ahh. yes14:39
mordredfungi: you know, it might also be time to split a few of those vhosts into their own hosts14:40
sdaguehashar: no, it's just sanity checking that it looks like a number14:40
*** gabriel-bezerra has joined #openstack-infra14:40
fungimordred: that dawned on me after i started seeing other things like loglevel parameters landing in the error log too. probably something i should point out to clarkb/jhesketh once they're around14:40
hasharsdague: so that is ElasticSearch consuming Gerrit stream-events and handling the bug number right?14:41
gabriel-bezerraHi folks! Are the bugs listed on ordered in any way? Importance, most frequent, oldest first???14:41
mordredgabriel-bezerra: most important/frequent first14:41
sdaguehashar: nothing is sanity checking the value14:41
sdaguerecheck bug 000000000000000000000000000000000000000000000000000014:41
mordredgabriel-bezerra: or, rather, most frequent, which implies important14:41
sdaguewill trigger things14:41
mordredsdague: ssssh. dont let the devs knows that ...14:41
gabriel-bezerramordred: Good. Thank you.14:42
sdaguemordred: they already do14:42
mordredsdague: damn14:42
hasharsdague: but then the bug number is never passed as a parameter to the jenkins jobs right?14:42
sdagueI've seen: recheck bug 12345678914:42
sdaguehashar: correct14:42
hasharI could not find anything about it in Zuul but wanted a confirmation14:42
hashargreat :-D14:42
fungimordred: and yeah, we coalesced separate servers onto the static.o.o host but yeah, the pypi site and possibly the logs site are getting to the point where they might benefit from being separated again14:42
sdaguehashar: mostly we're trying to use it as provided info so we can datamine the issues that we don't know about yet14:42
mordredfungi: we could also consider lb-farms too14:43
openstackgerritMatt Riedemann proposed a change to openstack-infra/elastic-recheck: Add query for nova unit test bug 1339739
uvirtbotLaunchpad bug 1339739 in nova "test_cidr_repr race fail with "AttributeError: 'PostgreSQLOpportunisticFixture' object has no attribute '_details'"" [Undecided,New]
mordredfungi: especially for "serve logs from swift" - that's like one of the classic reasons to have an elastic loadbalanced farm of web servers...14:43
fungimordred: yep, in fact we already do that in one place, so could probably scale that design to additional backends if we wanted (though making it more of a high availability config and doing a pair of load balancers could help too)14:44
mordredJayF: done14:44
JayFmordred: thank you very much14:44
*** rkukura has joined #openstack-infra14:44
fungimordred: or try to take advantage of load-balancing services offered by our providers, though those would be provider-specific things so meh14:45
fungigiven that we want openstack to support consistent services across providers, i'd rather avoid leveraging provider-specific features where possible14:46
fungifeels sort of like we'd be endorsing those differences14:46
*** morganfainberg_Z is now known as morganfainberg14:46
mordredfungi: I totally agree14:46
*** obondarev has joined #openstack-infra14:48
*** gabriel-bezerra has left #openstack-infra14:48
*** dprince has joined #openstack-infra14:48
openstackgerritA change was merged to openstack-infra/config: Add IPA CoreOS Imagebuilder to check + gate
*** thedodd has joined #openstack-infra14:49
*** zz_gondoi is now known as gondoi14:53
mordredSergeyLukjanov: ooh. mod_proxy_uwsgi exists in trusty14:53
mordredthat means as soon as we can run infra things on trusty, we can move stackalytics to that and then I think we'll be set14:54
*** dpyzhov has joined #openstack-infra14:54
*** doddstack has joined #openstack-infra14:54
*** thedodd has quit IRC14:57
rgerganovjeblair, thanks for the review on
openstackgerritA change was merged to openstack-infra/elastic-recheck: Add query for nova unit test bug 1339739
uvirtbotLaunchpad bug 1339739 in nova "test_cidr_repr race fail with "AttributeError: 'PostgreSQLOpportunisticFixture' object has no attribute '_details'"" [Undecided,New]
rgerganovjeblair, does changing "ensure => link" with "ensure => present" will suffice?14:58
rgerganovsorry, I am not really a puppet guy14:59
jeblairrgerganov: maybe?  i don't know off hand.  i can test it though14:59
*** datsun180b has joined #openstack-infra14:59
rgerganovjeblair, ok, I will update the patch14:59
*** Sukhdev has joined #openstack-infra15:00
openstackgerritMatt Riedemann proposed a change to openstack-infra/elastic-recheck: Categorize fixed fail for nova bug 1330431
uvirtbotLaunchpad bug 1330431 in nova "wrong lock name when operating instance events" [Medium,Fix committed]
*** HenryG has joined #openstack-infra15:02
*** YorikSar has joined #openstack-infra15:02
openstackgerritRadoslav Gerganov proposed a change to openstack-infra/config: Add button that shows/hides CI comments in Gerrit
*** weshay has quit IRC15:02
fungimordred: deeper analysis of that socket timeout shows that pip did manage to retrieve cryptography from the mirror, then started digging CPython and cffi off the mirror to build it. apache did return cffi with a 200/ok but the opaqueness of the easy_install output doesn't say if it timed out retrieving that or something it wanted after cffi15:04
*** afazekas has quit IRC15:04
fungii'll compare with apache logs corresponding to a successful run and see if i can suss out what it would have wanted next15:04
jeblairfungi: do the errors tend to come from easy_install?15:04
fungibut so far no evidence of errors corresponding to that15:04
fungijeblair: possibly--i need to do a broader analysis now that i know what i'm looking for15:05
jeblairrgerganov: nope, that made a 0 byte file15:05
fungijeblair: though if that turns out to be the case, no amount of new retry features in pip 1.6 will help us there15:05
fungithough the setuptools replacement within pip might15:05
rgerganovjeblair, ok, I think I need to use "source" instead of "target" when it is not a linik15:06
*** hashar has quit IRC15:06
jeblairrgerganov: yes!15:06
rgerganovjeblair, update is coming15:06
fungibut not sure what dstufft's eta is on setuptools/easy-install reimplementation within pip15:06
*** sabeen1 has joined #openstack-infra15:06
openstackgerritRadoslav Gerganov proposed a change to openstack-infra/config: Add button that shows/hides CI comments in Gerrit
jeblairrgerganov: that did it15:08
jeblairi'm restarting gerrit on review-dev now just to be certain15:08
rgerganovjeblair, great!15:08
jeblairshould demonstrate it15:09
*** zhiyan is now known as zhiyan_15:10
rgerganovyes, it works the same way as with the greasemonkey script15:10
rgerganovwhich is expected :)15:10
SergeyLukjanovmordred, cool15:10
*** _nadya_ has quit IRC15:10
*** hogepodge has quit IRC15:11
*** reed has joined #openstack-infra15:12
*** baohua has quit IRC15:12
openstackgerritMatt Riedemann proposed a change to openstack-infra/elastic-recheck: Categorize horizon gate rebase unit test fail
*** rdopieralski has quit IRC15:12
*** jlibosva has quit IRC15:13
*** lcheng has quit IRC15:14
*** hogepodge has joined #openstack-infra15:15
fungimordred: so it looks like these just started cropping up in the past 48 hours...15:15
fungimordred: all hits are on hpcloud nodes, none on rax. could be coincidence, could be something a local mirror in hpcloud would help15:16
openstackgerritA change was merged to openstack-infra/elastic-recheck: Categorize fixed fail for nova bug 1330431
uvirtbotLaunchpad bug 1330431 in nova "wrong lock name when operating instance events" [Medium,Fix committed]
*** lcheng has joined #openstack-infra15:17
jeblairfungi, mordred: i've been thinking about it, and i'm still a little worried about running an hpcloud mirror.  basically: what happens when hpcloud deletes it?15:17
jeblairof course, we have to weigh the risk that they might delete it, on average, say, every 6 months against the fact that we have sporadic errors frequently now15:18
jeblairand maybe we say the risk is worth it15:18
jeblairbut still, when that happens, we'll fast-fail all of the tests that run on hpcloud15:18
openstackgerritNikita Konovalov proposed a change to openstack-infra/storyboard: Update _route method usage
*** che-arne has quit IRC15:18
*** afazekas has joined #openstack-infra15:18
fungijeblair: mordred: is that still their modus operandi or have they changed things lately?15:19
jeblairfungi: i don't know; i haven't been a contact on an hpcloud 1.1 account yet, so i haven't gotten any "too bad" emails15:19
*** SpamapS_ is now known as SpamapS15:21
*** SpamapS has quit IRC15:21
*** SpamapS has joined #openstack-infra15:21
*** sballe has quit IRC15:21
sdaguefungi: I think - message:"Complete output from command python egg_info" AND tags:console is a better match15:21
openstackgerritA change was merged to openstack-infra/elastic-recheck: Categorize horizon gate rebase unit test fail
sdaguethat shows about 10x the hits of the one you posted15:21
*** jaypipes is now known as jay-house-hunter15:22
sdaguethough that does include refstack false positives15:22
*** tsg has joined #openstack-infra15:25
*** fanhe has quit IRC15:25
fungisdague: possibly. so far the socket timeouts i'm finding from the pattern i searched for are all bubbling up from easy_install15:26
sdagueyeh, too bad we don't collapse tracebacks in console like we do in service logs15:28
sdaguethat would make it easier to match15:28
*** hashar has joined #openstack-infra15:30
*** jgallard has joined #openstack-infra15:30
fungibut there certainly could be other manifestations of whatever the underlying issue is15:31
fungiso far i'm not finding any evidence that this is due to increased load on the pypi mirror15:32
openstackgerritMonty Taylor proposed a change to openstack-infra/config: Run stackalytics in infra
mordredjeblair: my irc bouncer has been running in hpcloud for over a year and I've gotten no too-bad emails15:33
*** david-lyle has joined #openstack-infra15:33
openstackgerritNikita Konovalov proposed a change to openstack-infra/storyboard: Team REST API added
*** ihrachyshka has quit IRC15:34
*** lttrl has joined #openstack-infra15:34
jeblairmordred: that's promising15:34
mordredjeblair: also, btw ^^ the patch I just sent in should work for running stackalytics via uwsgi - but it requires trusty15:35
mordredjeblair: just fyi15:35
pelixzaro: clarkb has identified a potential issue with the cmd tests for JJB. When adding a test (see on top of the initial ones you created, it triggered a failure due to CacheStorage attempting to create the same directory in parallel for multiple tests.15:36
fungimordred: in fact, the cacti graphs don't suggest there's any significant increase in traffic volume (bandwidth utilization, packet count, tcp open/established) either15:36
mordredfungi: stellar15:36
*** Longgeek has quit IRC15:36
*** masayukig has quit IRC15:37
fungimordred: so i suspect that all the high-download-volume offenders were already restricted to the mirror, and the infra/stackforge projects which are now also downloading everything from teh mirror were a drop in the bucket by comparison15:37
*** Longgeek has joined #openstack-infra15:37
*** asettle has joined #openstack-infra15:37
zaropelix: will take a look15:39
*** pballand has joined #openstack-infra15:39
*** hogepodge has quit IRC15:39
pelixzaro: I've put the two suggestions into the review of the cmd tests
pelixafternoon to you :)15:40
openstackgerritSean Dague proposed a change to openstack-infra/elastic-recheck: change to correct quoting, ER doesn't seem to pick this up
*** Longgeek has quit IRC15:40
*** Longgeek has joined #openstack-infra15:40
joa hello, I got a question about jenkins configuration for third-party CI (got most of the thing set up, only wondering about how to go for some details)15:40
*** msabramo has quit IRC15:41
*** Guest58876 has quit IRC15:41
*** moted has joined #openstack-infra15:42
*** masayukig has joined #openstack-infra15:42
joa-> About the comments triggers ( 'recheck no bug' and 'recheck bug ####' ), since I'm using the gerrit and gerrit-trigger plugins; I think it's related to adding a trigger on 'comment added' event; But I'm not sure what value to check for, nor how to actually manage (and maybe display the related result?) it, since there's a "Verdict" category within the configuration interfavce.15:42
*** che-arne has joined #openstack-infra15:43
*** hogepodge has joined #openstack-infra15:44
*** elambert has joined #openstack-infra15:44
fungijoa: openstack-infra uses something called "zuul" instead of the gerrit-trigger plugin (it was actually our replacement for it since it was inadequate for our needs). as a result we don't have a lot of current expertise in that plugin... though i suppose there might be some people lurking in here who know how to configure it15:45
*** alexpilotti has quit IRC15:45
*** jgallard has quit IRC15:45
joaWell, if you can advise me on how to go about using zuul and configure it properly, I'm not against it :p15:46
*** marun has joined #openstack-infra15:46
*** lttrl has quit IRC15:46
joaoriginally, I was trying to follow two pieces of info: jaypipe's article with his collections of scripts (to setup jenkins master and slave); and
fungijoa: though i think jaypipes may have written some blog posts on using zuul for third-party ci as well15:46
fungiand yeah, i think we probably need to remove the gerrit-trigger section from third_party.html since it gives the impression that we can answer questions about it when people have trouble using it15:48
joa :p15:48
*** otherwiseguy has joined #openstack-infra15:48
joaMhh I need to find back again jaypipe's article15:49
joakind of lost the link recently15:49
*** _nadya_ has joined #openstack-infra15:50
datsun180bso i found but i cannot seem to find the base endpoint that would recieve my requests, and it doesn't appear to be review.o.o. what am i going wrong here?15:50
fungidatsun180b: it is review.o.o... did you add an http password to your account settings and authenticate your requests with it?15:51
*** msabramo has joined #openstack-infra15:51
datsun180bi hadn't, i was just probing and figured at least some of it would be public, even if all i got was a catalog15:52
fungidatsun180b: i *think* all the api calls require authentication15:52
openstackgerritA change was merged to openstack-infra/elastic-recheck: change to correct quoting, ER doesn't seem to pick this up
jeblairdatsun180b: they don't all require authentication, but the list of endpoints is right there in the doc you linked15:53
datsun180bbut thanks, i'll work from there15:53
*** hashar is now known as hasharMeeting15:53
joaSeems like I might (once again) have missed a step with jaypipes'scripts; no job appear to be preconfigured after starting zuul's service.15:54
fungijoa: do the jobs exist on your jenkins master? if so, you might have missed installing the jenkins-gearman plugin, or missed restarting your jenkins master after installing it, or missed configuring it15:55
fungior you might have a firewall between your jenkins master and zuul server blocking gearman traffic15:55
openstackgerritKhai Do proposed a change to openstack-infra/jenkins-job-builder: tests for JJB command processing
zaropelix: ^15:55
joafungi: Well the jenkins master has been restarted a few times since install; but no jobs did appear. Also, it seems like I forgot to start zuul services. So I'm trying to redo jenkins jobs update + zuul restarting15:56
openstackgerritDarragh Bailey proposed a change to openstack-infra/jenkins-job-builder: Allow `test` command to use conf files
marunclarkb, sdague, fungi: I switched to using rootwrap in the neutron functional job instead of raw sudo but I'm still seeing errors related to the command not being password-less.15:57
pelixzaro: reviewed :)15:57
*** amcrn has quit IRC15:57
marunsearch for RuntimeError in
fungidatsun180b: yeah, the projects list for example is accessible without authentication. i get json output from wget -O- ''15:57
marunis there something I'm missing?15:57
*** afazekas has quit IRC15:58
pelixzaro: I preferred the mock approach myself as well since it avoids the test code needing to know how CacheStorage is implemented15:58
datsun180bfungi: cool, thanks. all i'm really trying to do is autobuild review dependency trees and i could either do it by hand or make robots use the api to do it15:59
datsun180bthere isn't a tree-style view of open reviews already is there?15:59
fungimarun: you might want to add a set -x in your post-test hook. the log doesn't have enough tracing to pinpoint the problem easily...
anteayafungi: sorry I didn't know that the oslo gerrit groups renaming should have taken place prior to the merging of the olso acl uniformity patch, I will try to remember to get the order correct next time16:00
fungianteaya: it wasn't a huge deal, it just ends up leaving empty hidden groups behind as cruft since gerrit doesn't have a group deletion option (or didn't used to have one, though maybe they added it in a more recent release)16:01
joaWell.. No steps tell how/what to configure for zuul, and no job appear after the restarts.. Guess I'll have to dig deeper16:01
fungijoa: is where we publish zuul documentation16:02
anteayafungi: ah okay, well I will work in the direction of lack of cruft16:02
anteayafungi: thanks16:02
fungijoa: we also have some documentation about how we configure our zuul server for upstream testing, if it helps...
joaWell, the zuul logs do talk about some triggers it seems, but nothing on jenkins side16:03
joaI'll go through that doc, fungi; thanks16:03
fungijoa: we've also got an overview slide presentation about our infrastructure, as well as a zuul-specific one, which may help provide some high-level/conceptual understanding...
joaoh okay, good for internal communication on my side, that will help :D16:05
*** markmc has quit IRC16:05
*** jay-house-hunter has quit IRC16:05
*** afazekas has joined #openstack-infra16:06
*** pballand has quit IRC16:06
datsun180boh my problem was i lacked curl-ca-bundle, https://review.o.o/projects worked just fine after installing that16:06
*** pcm__ has joined #openstack-infra16:08
*** gyee has quit IRC16:08
*** Evgeny_ has joined #openstack-infra16:09
*** pcm_ has quit IRC16:10
*** arnaud has joined #openstack-infra16:10
openstackgerritMonty Taylor proposed a change to openstack-infra/config: Run stackalytics in infra
*** asettle has quit IRC16:16
mordredfungi, jeblair, SergeyLukjanov: I was wrong! it's in precise too16:16
*** denis_makogon_ has joined #openstack-infra16:17
fungidatsun180b: aha, it was failing to validate the server certificate and not telling you. that's an awesome failure mode16:17
datsun180byeah, curl was saying exactly "Not Found" and giving me no other reasons16:18
*** chandankumar has quit IRC16:18
fungiwow. perl programmers... ;)16:18
*** nati_ueno has joined #openstack-infra16:19
*** asettle has joined #openstack-infra16:19
*** gyee has joined #openstack-infra16:22
*** reed has quit IRC16:22
*** arnaud has quit IRC16:23
*** e0ne has quit IRC16:25
*** jpich has quit IRC16:25
*** miqui has quit IRC16:27
*** miqui has joined #openstack-infra16:28
*** _nadya_ has quit IRC16:29
*** _nadya_ has joined #openstack-infra16:30
*** marcoemorais has joined #openstack-infra16:31
*** amcrn has joined #openstack-infra16:31
*** reed has joined #openstack-infra16:31
*** homeless_ has joined #openstack-infra16:31
*** whoops has joined #openstack-infra16:32
*** homeless has quit IRC16:34
*** _nadya_ has quit IRC16:34
fungiso apparently when the timeouts happen in pip instead of setuptools, it looks like
JayFIs there a way with git-review to push a change as WIP from the start?16:36
*** marcoemorais has quit IRC16:36
*** marcoemorais has joined #openstack-infra16:37
*** jistr has quit IRC16:37
openstackgerritMichael Krotscheck proposed a change to openstack-infra/config: Modularized StoryBoard Module
anteayaJayF: not that I know of16:37
juzuluagHi there. Trying to push a spec to got the following message:16:38
juzuluaggit review16:38
juzuluagfatal: ICLA contributor agreement requires current contact information.16:38
juzuluagPlease review your contact information:16:38
juzuluagfatal: The remote end hung up unexpectedly16:38
juzuluagThen, went to update the contact information, but it doesn't allow it to do it.16:38
anteayaJayF: so I'm putting my money on no16:38
fungiJayF: no, since wip is an openstack-gerrit-specific configuration we're avoiding encoding openstackisms in it directly. we've discussed the possibility of adding a feature which can set arbitrary label votes based on configuration it finds in the .gitreview file, possibly even overloading the -D (draft) option to do so16:38
anteayajuzuluag: please use paste.openstack.org16:38
*** marcoemorais1 has joined #openstack-infra16:38
JayFfungi: anteaya: Thanks, that's useful to know :)16:38
fungijuzuluag: were you following the instructions at ?16:38
anteayajuzuluag: yes it does, you update then click submit and the timestamp is updated16:38
*** marcoemorais1 has quit IRC16:39
*** jcoufal has quit IRC16:39
*** marcoemorais1 has joined #openstack-infra16:39
fungijuzuluag: specifically did you join the foundation by going to and make sure to use the same primary e-mail address there as the preferred e-mail address in your gerrit contact information16:39
juzuluagHi fungi. Yes I did join the foundation16:40
anteayadid you update your contact information and the submit it?16:40
anteayajuzuluag: did the timestamp update?16:40
fungijuzuluag: and your primary e-mail address at matches your preferred e-mail address at ?16:41
juzuluagIt didn't. I got an error: Code Revew Error. Cannot store contact information16:41
*** pballand has joined #openstack-infra16:42
*** marcoemorais has quit IRC16:42
*** jgrimm has quit IRC16:42
fungijuzuluag: that error message most often happens because of an e-mail address mismatch. please double check the primary/preferred e-mail addresses at the links i mentioned16:42
juzuluag'll do. Thanks.16:42
*** GheRiver1 is now known as GheRivero16:43
*** dizquierdo has quit IRC16:43
*** trinaths has joined #openstack-infra16:44
*** luqas has quit IRC16:46
clarkbpelix: sorry about that. I ddin't realize there was another change until after I made my comment. But looks like it got sorted anyways16:46
*** [1]trinaths has joined #openstack-infra16:47
clarkbfungi: log levels?16:47
anteayaSergeyLukjanov: I have a question on
anteayaSergeyLukjanov: line 16, it isn't part of the change, but I wondering if we need it16:47
fungiclarkb: look at /var/log/apache2/error.log on static.o.o16:47
anteayaSergeyLukjanov: my understanding is that granting -1/+1 to Registered Users is a permission on all projects16:48
anteayaSergeyLukjanov: does it need to be specified for stable branches?16:48
*** trinaths has quit IRC16:49
*** [1]trinaths is now known as trinaths16:49
fungianteaya: yes, because in the all-projects acl we have exclusive group permissions on code review for refs/stable/* so that the stable-maint team can be the +2 on those rather than the per-project core teams16:49
fungianteaya: so if that exclusive group permission is overridden in a project's acl, it has to re-add the ability for registered users to vote +/-116:50
clarkbfungi: the keystone logger thing is a python logging stderr output so it ends up in apache error log16:50
clarkbfungi: it is arguably a bug in keystoneclient that it doesn't start cleanly by default16:50
fungiclarkb: yeah, i was more wondering where it was coming from. os-log-analyze?16:50
*** hasharMeeting has quit IRC16:51
anteayafungi: thank you!16:51
clarkbyes, keystoneclient is used by swiftclient to do the osloganalyze to swift lookups16:51
fungiclarkb: okay, that confirmed my suspicions as to its source. thanks16:51
JayFIs there a way to get to the workspace of a gate job after it's already run?16:52
*** sweston has joined #openstack-infra16:52
*** marcoemorais1 has quit IRC16:52
*** dkliban_afk is now known as dkliban_lunch16:52
*** marcoemorais has joined #openstack-infra16:52
fungiJayF: only if we hold the worker after the job is assigned to it but before it completes (otherwise nodepool immediately recycles it)16:53
anteayafungi: were do I find the all-projects acl - was looking for something called project.config in openstack/ and can't seem to find anything16:53
JayFfungi: okay, thanks16:53
*** sweston_ has joined #openstack-infra16:53
*** sweston__ has joined #openstack-infra16:53
zaroanteaya: what do you mean?16:54
zaroanteaya: on the UI?16:54
*** miqui_ has joined #openstack-infra16:54
*** tkelsey has quit IRC16:54
fungianteaya: it's in the refs/meta branch-like ref in the All-Projects project16:54
clarkbanteaya: in gerrit at,access you can gitweb it or fetch it too16:54
anteayazaro: we have permissions on all projects16:54
anteayaoh access it via the gui, I was thinking there was a .config file somewhere16:54
*** miqui has quit IRC16:55
*** dpyzhov has quit IRC16:55
clarkbdoes need to be unsubscribed from openstack-dev?16:55
anteayaclarkb: I'd get behind yes16:55
zaroanteaya: you need to check it out.16:55
clarkbanteaya: there is but only in gerrit. we don't manage it with jeepyb as there is a chicken and egg where you need to have a sufficient all projects config in order to have jeepyb manage it16:56
anteayaclarkb: ah16:56
anteayazaro: what command checks it out for me?16:56
*** derekh_ has quit IRC16:56
*** Guest58876 has joined #openstack-infra16:56
zarogit clone ssh://user@localhost:29418/All-Projects; cd All-Projects; git fetch origin refs/meta/config; git checkout FETCH_HEAD16:57
anteayazaro: thank you16:57
*** praneshp has joined #openstack-infra16:59
fungianteaya: for a more detailed example16:59
*** amitpp has joined #openstack-infra16:59
*** vhoward has joined #openstack-infra16:59
fungi(obviously substitute your username for mine)16:59
*** ociuhandu has quit IRC16:59
anteayafungi: thanks16:59
pelixclarkb: no worries, very useful review in any case :)16:59
*** pcm__ has quit IRC17:00
*** Sukhdev has quit IRC17:00
fungianteaya: there's also this if you want to browse it...;a=tree17:01
*** denis_makogon has quit IRC17:01
*** denis_makogon_ is now known as denis_makogon17:02
*** dmakogon_ has joined #openstack-infra17:02
*** crc32 has joined #openstack-infra17:03
*** harlowja_away is now known as harlowja17:03
zarofungi: ohh cool, didn't know that was available on gitweb17:04
anteayafungi: awesome, thanks, I never knew about that view17:04
anteayatoday is obviously a great learning day, thank you17:04
*** pcm_ has joined #openstack-infra17:04
clarkbanteaya: zaro: if you need links like that in the future gerrit has a handy link on the gui acl page17:05
*** marcoemorais has quit IRC17:05
clarkbmakes it easy to find17:05
*** marcoemorais has joined #openstack-infra17:05
fungizaro: yeah, if you're browsing the acl in its management view (which for reasons i don't quite grok gerrit wants to restrict parts of based on your group membership) there's a "gitweb" link at the top of the page which takes you to the enture thing17:05
*** marcoemorais has quit IRC17:05
*** marcoemorais has joined #openstack-infra17:05
*** Ajaeger1 has joined #openstack-infra17:06
*** marcoemorais has quit IRC17:06
fungiwhich completely shatters whatever the viewing restrictions there are imposed on the acl in management view17:06
*** marcoemorais has joined #openstack-infra17:06
zaroohh yeah. i see it now. thanks.17:07
*** marun has quit IRC17:07
fungilike elsewhere in gerrit, role-based access restrictions seem to be badly implemented17:07
fungireally, if i were trying to lock down a gerrit from a visibility standpoint, the first thing i'd do is completely disable its gitweb feature17:08
*** mwagner_lap has quit IRC17:08
fungibut even then, you can pull acl files via the ssh api as well, and that's even harder to restrict17:08
clarkbfungi: you can restrict it with acls :)17:08
clarkbrefs/meta/* read block registered users17:09
fungioh, good point, you can set an access policy around refs/meta17:09
*** marun has joined #openstack-infra17:09
fungiyeah, that17:09
*** jedimike has quit IRC17:09
clarkbthough that may break because registered users will include your admins17:09
clarkbnot sure if you can override a block with a more specific group but doing that in general is possible17:09
fungiclarkb: probably exclusive group read access on it instead17:10
*** e0ne has joined #openstack-infra17:10
dhellmanndoes anyone have admin access on the -dev list? we've got a spam bot or something posting messages17:10
clarkbdhellmann: ttx and reed maybe? I guess technically we probably have a way of backdooring mailman17:10
dhellmannthanks, clarkb17:10
reedI do17:11
clarkbttx: reed ^ any chance you can deal with that?17:11
fungidhellmann: i can probably dig up a server admin account, but yeah ttx and reed are listed as the mods for it17:11
clarkbreed: great thank you17:11
fungiahh, reed is around. awesome17:11
reeddhellmann, there must be some mechanical turk-based attack going on17:11
dhellmannreed: have you seen the messages from ""?17:11
reeddhellmann, we've been dealing with spammers on askbot for over a week17:12
reedthey're humans, registering accounts and then spamming with slow bots17:12
reednot easily caught by rate limiting17:12
dhellmannreed: Ah, yeah, I saw some messages about that, too.17:12
fungithe dts68x messages look like probes trying to determine whether they can post ads or files (maybe malware or warez) to the list archives and link to them17:13
reedwhat's the best way to block this moron?17:13
reedunsub is not going to be effective17:13
dhellmannyou can tell mailman to not allow that address to post or to resubscribe17:14
dhellmannI don't know how long that will be effective.17:14
reeddhellmann, where should I do that? sorry, mailman still confuses me17:14
*** Evgeny_ has quit IRC17:15
dhellmannreed: let me see if I can find the page on one of the lists I maintain17:15
reedI set that imbecil to moderation, I'll block the messages once they pop up in the queue17:15
fungilooks like the spam is coming from some site called we should probably just block that completely17:16
*** marun has quit IRC17:16
anteayareed: can you delete prior posts?17:16
reedanteaya, uhm ... now you really ask too much :) I don't know how to do that17:16
reedit's funny, I thought that those were legitimate posts at first17:17
reednow I see that he posted something like test messages17:17
*** johnthetubaguy is now known as zz_johnthetubagu17:17
reedare you sure this is a spammer?17:17
Ajaeger1reed, you should be able to set the moderator bit on the member - that way you can review what he posts17:18
fungiyeah, i suspect it's a bot which someone hooked up to a mailbox they subscribed, and it replies to messages with its archive probes17:18
reedAjaeger1, thanks, that's the first thing I did17:18
dhellmannreed: under the "privacy options" page there is a "spam filters" page where you can add rules to automate the filtering; which will help until they pick another email address17:19
fungioh, though this one looks like it's an actual list reply...
Ajaeger1dhellmann: you beat me ;)17:19
reeddhellmann, thanks17:19
Ajaeger1reed "List of addresses which are banned from membership in this mailing list"17:19
fungibut subsequent replies from the same account do not look legit... so maybe it's just someone who's really bad at this e-mail thing17:19
dhellmannAjaeger1: you might have a newer version of mailman than I have, that sounds like an even better way to do it.17:20
reedfungi, I have the same suspicion17:20
reedI think we have spamassassin on our mail server so regular spam should not pass easily but I'm not sure about anything17:21
*** marun has joined #openstack-infra17:21
reedhonestly, mailman is quite in 'maintenance mode'17:21
*** Ryan_Lane has joined #openstack-infra17:21
reedmeaning that if things break there we need to pull experts in from somewhere else, there is no dedicated resource for it17:21
Ajaeger1dhellmann: mailman 2.1.17 for me17:22
fungihere's an ml reply from the same account back in may, so probably is actually someone having mua issues...
*** afazekas has quit IRC17:23
fungireed: if push comes to shove, i can dig into fixing mailman issues. i've both run mailing lists and mailman servers themselves in years gone by. it would be unfortunate to need to do so, but it's an option17:23
*** gangil_ has joined #openstack-infra17:23
gangil_Hey, I am running a third party CI for openstack which is supposed to run temptests, however it fails towards the end with the following error message
*** mmaglana has joined #openstack-infra17:24
reedfungi, indeed, that's what I meant :)17:25
reedfungi, also there are a couple of pending tasks for our mailman that never got completed17:25
fungireed: yep, saw you unassigning those from rockstar17:25
reedlike installing other language packs, migrating the old archives etc17:26
clarkbgangil_: is the relevant section (want testrepository release so bad)17:26
reedtasks that keep slipping off the radar17:26
*** melwitt has joined #openstack-infra17:26
clarkbgangil_: there are import errors in those files. If you install testrepository from source the output is much eaiser to read, but no release has been made with those changes17:26
reedspeaking of radars, I have a near collision now :) need to focus and complete my presentation for tonight17:26
*** doddstack has quit IRC17:26
*** abhishek___ has joined #openstack-infra17:27
Ajaeger1fungi, annegentle just started collecting requirements for - see . We'll send a summary but if you have questions, feel free to respond17:27
*** nibalizer has joined #openstack-infra17:28
gangil_clarkb: so you want me to install python-testrepository package on that machine to get a better idea of the error? just making sure I understand you correctly17:29
*** homeless_ has quit IRC17:29
*** homeless has joined #openstack-infra17:29
fungiAjaeger1: yep, the phone based support is probably going to be the hardest requirement to meet. that doesn't sound like a feature we could reasonably provide for infra-managed systems without contracting a third party provider to do that part for you17:29
clarkbgangil_: you don't need to. I pasted the important parts of the error for you already17:30
abhishek___Hi all, can any one review the patch, and help me to get this merged17:30
*** yaguang has quit IRC17:30
gangil_clarkb: oh ok. cool I will take a look17:30
fungiabhishek___: this is the channel we use for managing openstack project infrastructure. you probably want to bring your change to the attention of the requirements core reviewers instead,members17:31
*** Evgeny has joined #openstack-infra17:31
Ajaeger1fungi: feel free to discuss this. I would be fine by infra's great IRC support if you find a superior solution going forward ;)17:32
openstackgerritJoe Gordon proposed a change to openstack-infra/elastic-recheck: Add query for bug 1326998
abhishek___fungi: thank you17:32
uvirtbotLaunchpad bug 1326998 in nova "UnboundLocalError: local variable 'domain' referenced before assignment" [Medium,Fix committed]
*** e0ne has quit IRC17:32
*** elambert has quit IRC17:33
clarkbAjaeger1: fungi: I think there is definitely an argument to be made that phone based support isn't necessary17:33
clarkbespecially considering the times we have run up against it17:33
fungiAjaeger1: yep, will do. so far we (the infrastructure team) doesn't have "phone support" for anything. i'm slightly curious when it's come in handy with rackspace and hasn't been needed completely because of something they broke17:33
clarkbfungi: those are the only times I have been aware of17:34
fungiin general "you can call us every time we break something and ask us to fix it again" doesn't seem like much of a feature17:34
*** arnaud__ has joined #openstack-infra17:34
*** e0ne has joined #openstack-infra17:34
fungii'd rather see us let the docs team and the rest of the project carefully review what's changing and then just not unexpectedly break things17:35
*** blamar has quit IRC17:35
*** mwagner_lap has joined #openstack-infra17:36
fungias for the service providers who provide the servers to us, they do actually have phone numbers we can call, but it's generally faster to open a case through their management portal or just ping their employees in irc17:36
Ajaeger1fungi, clarkb: Looking at the list, I see there are no *features* we need - we need a reliable service that scales great.17:37
jeblairthat we can do :)17:37
Ajaeger1So, no limitations for a change in that regard ;)17:37
fungithat sounds like something we can definitely provide, yes17:37
*** lcheng has quit IRC17:37
Ajaeger1jeblair, fungi: Indeed. It's serving static pages without interruption17:37
*** Sukhdev has joined #openstack-infra17:38
jeblairalso, we can probably refactor requirement #1 into "ability to recreate/restore old content"; the .snapshot directory and retention period are artifacts of the current implementation17:38
*** jerryz has joined #openstack-infra17:38
jeblairi think the ability to revert, etc, is probably the main thing that's getting at17:39
annegentlefungi:  clarkb Ajaeger1: Ha, I nearly struck out the "phone-based" since I really just need "some channel of communication"17:39
clarkbjeblair: and completely unnecessary in our git based world assuming maven nexus doesn't delete deps to build arbitrary versions17:39
*** zzelle__ has joined #openstack-infra17:39
*** afazekas has joined #openstack-infra17:39
fungiannegentle: we maintain lots of channels of communication ;)17:39
annegentlefungi: yep :)17:39
jeblairlike 50 :)17:39
*** zzelle__ has quit IRC17:39
harlowjafungi jeblair so i know we had the DMARC/DKIM discussion a while ago, can I bring it up again (or maybe later)17:40
annegentlejeblair: I've only had to use the .snapshot directory once, and I think it was for an .htaccess file problem, not a content file problem17:40
Ajaeger1clarkb: can we still build documentation from folsom branches if needed?17:40
clarkbAjaeger1: yes17:40
clarkbAjaeger1: again given the above assumption :)17:40
jeblairharlowja: i believe the current thinking is to ban domains that use dmarc from subscribing to the list because they cause innocent bystanders to be forcibly unsubscribed17:40
*** atiwari has joined #openstack-infra17:41
clarkbAjaeger1: but as long as we can grab a precise image (for java 6) and git clone on there I have a reasonably high expectation that that would work17:41
anteayagangil_: consider joining us at the weekly third party meeting:
Ajaeger1clarkb: good17:41
*** zzelle__ has joined #openstack-infra17:41
jeblairharlowja: i don't think we've actually done that yet; just waiting for someone to get suitably frustrated by repeated unsubscription to implement it17:41
openstackgerritA change was merged to openstack-infra/elastic-recheck: Add query for bug 1326998
uvirtbotLaunchpad bug 1326998 in nova "UnboundLocalError: local variable 'domain' referenced before assignment" [Medium,Fix committed]
gangil_anteaya: sounds good17:41
anteayagangil_: great17:42
openstackgerritMatt Riedemann proposed a change to openstack-infra/elastic-recheck: Add query for nova ec2 concurrency bug 1338841
uvirtbotLaunchpad bug 1338841 in nova "EC2KeysTest fails in tearDownClass with InvalidKeyPair.Duplicate" [High,New]
fungiharlowja: jeblair: the trick being to find a way to automatically identify those domains (we can preload it with the largest known offenders, is a big source of headaches there for example, and then simply react to complaints on smaller fry as they come up)17:42
harlowjajeblair me so sad, thats discriminating against a larger and larger population of the internet (not so  open) :(17:42
jeblairharlowja: actually, we're just helping yahoo implement the policy they have chosen.  it's a public service really.17:43
clarkbis that why I have been unsubscribed a couple times? or is it other gmail derp?17:43
fungiharlowja: there are two internets. one which corporate e-mail addresses use to send messages to each other, and one which developers use to communicate on technical mailing lists17:43
clarkbI should really do the sup17:43
jeblairharlowja: afaik, yahoo's policy on the matter is "employees should not use to subscribe to public mailing lists"17:43
jeblairharlowja: (and is the policy most dmarc domains have)17:43
*** davidlenwell_ is now known as davidlelnwell17:44
fungii personally have no qualms firewalling off the dmarc-using mail systems from those which chose not to implement them17:44
*** davidlelnwell is now known as davidlelnwell__17:44
*** davidlelnwell__ is now known as davidlelnwell17:44
*** davidlelnwell is now known as davidlenwell17:44
jeblairharlowja: please don't accuse us of discriminating; it's yahoo and other dmarc systems that are executing the denial of service here, and doing it by violating the email rfcs.17:45
fungiit makes the choices those entities have made much more obvious (and effective at their end goal of getting rid of e-mail entirely)17:45
harlowjawhen gmail turns it on, won't that be bad :-/17:45
jeblairharlowja: it's already bad17:46
*** dkliban_lunch is now known as dkliban17:46
fungigmail users will have fun sending messages back and forth to yahoo users17:46
harlowjaand also to the ML right?17:46
harlowja*since it will be blocked17:46
jeblairharlowja: i think fungi meant that gmail and yahoo will be on their own private email system then and only able to email each other :)17:47
fungiyes, that17:47
fungian exaggeration, admittedly17:47
SpamapSjeblair: is there some canonical document that explains this problem in detail? My curiosity is now piqued.17:48
harlowjahmmmm, so when that happens (already happened for *yahoo* emails) then those people will be unable to communicate effectively with the mailing list; isn't that bad and not the openness we want to have in the community?17:49
jeblairSpamapS: has some info17:49
*** shashankhegde has joined #openstack-infra17:49
openstackgerritA change was merged to openstack-infra/elastic-recheck: Add query for nova ec2 concurrency bug 1338841
uvirtbotLaunchpad bug 1338841 in nova "EC2KeysTest fails in tearDownClass with InvalidKeyPair.Duplicate" [High,New]
Ajaeger1I have a couple of translation reviews in the queue including these two that I'd like to see merged soonish -, - could you have a look, please?17:51
clarkbharlowja: forgive my rudimentary understanding of the problem, but mailman isn't doing anything to harm openness. Other entities like yahoo have chosen to not receive mail from our mailman17:51
* harlowja i'm not really trying to push one way or the other, just want to be open, and knowing that others are turning on more of DMARC and it affects the usage of the ML (and makes me sad)17:51
fungiharlowja: it's bad that they choose to rely on e-mail providers who want to restrict their ability to discuss things on public mailing lists. i think in general they want to push them to their own group discussion forums instead where they can more easily track and advertise at them17:51
clarkbharlowja: I don't understand how that is us not being open17:51
clarkbharlowja: that is yahoo self isolating itself17:51
SpamapSjeblair, fungi: thanks17:51
clarkbfungi: that17:52
anteayaharlowja: there is a difference between being open (which we are) and chasing around after people who are making non-open choices17:53
anteayaopen is a standard, and we are implementing it17:53
anteayaand welcome others who implement the same standard - open17:53
fungiarguably dmarc is also a standard, but it's a standard which limits a previous and very widely-used existing standard17:53
* mordred agrees that we want openness and it will be bad for everyone if more people continue to break the internet17:54
mordredhowever, there is not a lot we can do about more and more people breaking the internet17:55
mordredother than do what we can to limit knock-on effects17:55
mordredsuch as - if yahoo's decision to break the internet also breaks something for someone else who didn't break the internet via our systems17:55
harlowjaopen, not open, standard, not standard, good standard, bad standard.... people are applying it whether or not we all agree with it (i honestly don't care if its on/off), it just limits those who can use the ML (and that seems bad for people that are trying to be involved)17:55
mordredwe can do what we can to prevent yahoo's brokenness from affecting the other people17:55
fungimy main concern with dmarc is that it doesn't actually provide any direct benefit to the end user, as it doesn't certify/validate the endpoints of the communication channel (the people sending and receiving messages) and instead only identifies the communication channel between mailservers for the benefit of the mass mail providers17:55
mordredharlowja: yes. I agree17:55
*** asettle has quit IRC17:55
mordredI'm upset about the effects of it17:56
jeblairamusingly, the dmarc discussion list itself runs on mailman; last i checked, it has the same problems everyone else does.17:56
mordredharlowja: but as i said above, there is very little we can do about the decisions of some companies17:56
fungiyep, irony there17:56
*** weshay has joined #openstack-infra17:56
mordredall we can do is try to limit that damage to the people directly affected so that it doesn't become WORSE17:56
*** abhishek___ has quit IRC17:57
jeblairfungi: they decided _not_ to pursue ietf standardization17:57
fungijeblair: yeah, note i didn't say "ietf standard" ;)17:57
* harlowja redirects this from only being a yahoo thing (its not)17:58
jeblairwell, yahoo has been leading it17:58
harlowja ;)17:58
jeblairyahoo started dkim17:58
fungibut in general i think dmarc is a few large players in the e-mail commoditization industry trying to make things easier/more profitable for themselves, at the expense of an open standard which fewer and fewer people care about supporting/using any more (it's all sms and tweets and facebook walls with the young kids now)17:59
*** ildikov has joined #openstack-infra17:59
*** denis_makogon has quit IRC17:59
*** asettle has joined #openstack-infra17:59
jeblairand yahoo submitted the dmarc draft to the ietf before they dropped it17:59
*** mestery has quit IRC18:00
harlowjajeblair k, i didn't know that (i'm not involved in any of that community/people)18:00
mordredharlowja: we keep mentioning yahoo because you work there and we have a vague hope that you can effect positive change within your organization18:00
harlowjai know18:00
fungipreviously we (normal mail systems administrators) mostly just ignored dmarc and expected it to go away, but it didn't, and gained traction which is now going to be hard to reverse18:00
mordredI suppose I could call marissa18:00
mordredor have meg call her18:00
harlowjamordred lol18:00
mugsiedhellmann: you around? wanted to check when the full release of oslosphinx is....18:00
*** mestery has joined #openstack-infra18:00
*** Longgeek has quit IRC18:01
harlowjamordred i've been sending some of these logs to them (or some of them that i think might be these people), hoping they can read it over at least and undersetand, so maybe that helps18:01
harlowjacan try to get them in here, but it might just not be useful at that point (vi vs emacs battle ensues)18:02
*** Alexei_987 has joined #openstack-infra18:02
fungibut i definitely think implementing workarounds which make the negative impact of dmarc less obvious does nothing to help us get rid of it18:02
jeblairi expect that if we ban subscriptions from dmarc domains, we'll have some very nice text explaining why that people can send around18:03
*** msabramo has quit IRC18:03
harlowjajeblair, it will just be interesting when it bans gmail (talking to folks here that is just a matter of time)18:03
mordredok. I just poked sarob about it too18:03
harlowjamordred k, he sits next to me, i can ask him to get in here18:04
fungiand i think the one thing that the dmarc drafters failed to take into account, in knowingly breaking a valid if somewhat minority use case for e-mail, is that most of the e-mail systems administrators on the internet are actually part of that minority they've disenfranchised by breaking mailing list discussions18:05
clarkbharlowja: so far gmail has been good about making it clear they cannot verify source of email and so on. It also tries to spam filter your messages but I have an explicit rule that all ml mail is not spam18:05
*** shayneburgess has joined #openstack-infra18:05
*** Guest58876 has quit IRC18:05
*** jcoufal has joined #openstack-infra18:05
clarkbharlowja: it would be odd to me that they would go through all that effort just to shut it off. but its google so wouldn't surprise me18:05
*** sarob has joined #openstack-infra18:05
*** msabramo has joined #openstack-infra18:06
clarkbin any case this is why I have the other account. I just need a client that isn't terrible for mail lists18:06
harlowjasarob hi there!18:07
clarkbjeblair: this is when you tell me to use that emacs gnus thing :)18:07
jeblairclarkb: yeah that!18:07
fungiclarkb: or mutt. mutt's great18:07
nibalizermutt for lyfe18:08
clarkbI have actually been eyeing sup but lazyness has prevented me from setting it up18:08
jeblairsarob: so there's the dmarc thing yahoo did a while ago, and it's breaking mailing lists...18:08
*** atiwari has quit IRC18:08
*** amitpp has quit IRC18:09
*** tusharsg has joined #openstack-infra18:09
* jesusaurus just wishes we could leave email in the 80s and move on with our lives18:09
jeblairsarob: the upshot is that if people from dmarc domains subscribe to mailing lists, innocent bystanders get unsubscribed when they post to the list18:10
*** amitpp has joined #openstack-infra18:10
*** marcoemorais has quit IRC18:10
jeblair(innocent bystanders get unsubscribed when people from dmarc enforcing domains post to the list)18:10
*** marcoemorais has joined #openstack-infra18:11
*** xianghui has quit IRC18:11
*** marun has quit IRC18:11
dims_infra-cores, please bless oslo.utils when you get a chance (
*** marcoemorais has quit IRC18:11
*** amitpp has quit IRC18:11
*** marcoemorais has joined #openstack-infra18:11
*** marun has joined #openstack-infra18:11
sarobjeblair: so why can't the smtp gateways just be updated to support dmarc?18:12
jeblairsarob: from what i've heard the suggestion from dmarc-enforcing domains (like yahoo and twitter) to their employees is "use a personal email account to subscribe to mailing lists"18:12
*** tsg has quit IRC18:12
*** doude has quit IRC18:12
jeblairsarob: it's a fundamental design flaw in dmarc -- it is actually designed to break mailing list software (it is incompatible with the ietf email rfcs)18:13
sarobjeblair: but are not other public email domains moving towards supporting the same spam avoidance?18:13
fungisarob: the current workarounds suggested by for mailing lists providers are described at for a bit more background18:13
harlowjaas long as personal email is also not (the non-corporate version) ;)18:13
*** thedodd has joined #openstack-infra18:14
*** Evgeny_ has joined #openstack-infra18:14
sarobthis seems like a losing battle if the rest of the public email going the same way18:14
*** pcm_ has quit IRC18:15
jeblairsarob: it's not clear that's the case.  moreover, there are not really any other options right now18:15
*** hemna_ is now known as hemna18:15
dims_jeblair, did you see what the ASF folks did?
*** Evgeny has quit IRC18:16
*** Evgeny_ has quit IRC18:16
jeblairsarob: all of the workarounds have significant problems, and moreover, generally aren't implemented in free software mailing list systems18:17
*** pcm_ has joined #openstack-infra18:17
*** atiwari has joined #openstack-infra18:17
dhellmannhey, folks, I've had another question about getting requirements updated to use alpha versions of oslo libs. The bandersnatch cut-over was related to fixing that, right? Was there anything else we needed to land? It seems like there was one other patch...18:18
fungidhellmann: i think maybe there was --no-use-wheel in pip calls in some places (maybe in tox.ii files or something) which would have prevented it from working?18:19
jeblair(with the possible exception of #1, which i'd probably be okay with :)18:19
sarobjeblair: so dmarc support is not soon on the horizon?18:19
fungisarob: dmarc support is not really viable. dmarc seems to be designed to push group discussion off of e-mail to other fora18:20
dhellmannfungi: in infra/config or in projects? I can handle updating the projects, but wanted to make sure we expect it to work before I do something like recheck for example18:20
jeblairsarob: it's not an issue of "support" as much as deciding which of the ways of breaking mailing list software is least harmful, and then implementing "support" for that.  it's not like spf or dkim where it's just writing some code.  in the current spec, there's no way for list software to continue to operate as it has.18:21
fungijeblair: option #1 would prevent us from adding list unsubscribe information in message bodies, even as mime attachments, though i suppose it's possible we could live without that18:22
fungidhellmann: in individual projects i think. maybe in devstack too? i'm git grepping now18:23
clarkbI don't think individual projects do that18:23
*** trinaths has quit IRC18:23
clarkbfungi: dhellmann: instead they relied on (and possibly still do) the older pip behavior of not installing wheels18:23
*** dstanek is now known as dstanek_zzz18:23
dhellmannfungi: ok, don't spend any time on it, I'll recheck that patch and we can worry about the other projects when they have trouble updating18:24
jeblairthe official "we throw up our hands" response from yahoo about mailing lists18:24
fungiclarkb: dhellmann: yeah, i'm not finding any. it's possible it was just pypi-mirror which was doing that, and we don't use pypi-mirror now anway18:24
dhellmannfungi: cool18:24
*** bdpayne has joined #openstack-infra18:24
* dhellmann rechecks...18:24
*** dims_ has quit IRC18:24
*** markmcclain has joined #openstack-infra18:24
fungiyep, only obvious instance i'm finding is in openstack-infra/pypi-mirror:pypi_mirror/cmd/run_mirror.py18:25
openstackgerritGregory Haynes proposed a change to openstack-infra/pypi-mirror: Force --no-use-mirror installaions
*** morganfainberg is now known as morganfainberg_Z18:25
fungiso unless there was some other blocker, should be fine now18:25
*** dims_ has joined #openstack-infra18:25
*** pelix has quit IRC18:27
jeblairsarob: to distill a bit -- i think the issue is that the dmarc working group did not solve the "mailing list problem" and in fact mostly just ignored it; ideally before this is widely implemented, there should be a plan for accomodating mailing lists, and the dmarc group should work to implement the needed features in free software systems18:27
mordreddhellmann, fungi: we need to fix now18:28
sarobi suggest we move all mailing lists to twitter using hashtags18:28
* mordred throws a wet cat at sarob18:28
jeblairsarob: but instead, everything i've read on the subject indicates they just didn't care.18:28
clarkbmordred: adrian_otto pushed a start for that18:28
harlowjasarob does pine support twitter?18:28
mordredclarkb: oh, he did?18:28
* sarob get it off, get it off18:28
clarkbmordred: yup, because I mentioned that it was part of bandersnatch move and it affected requirements tests as they don't use the mirror but probably should18:29
fungione upside to switching to twitter is that nobody would write novel-length ml posts any longer ;)18:29
* harlowja including me, haha18:29
sarobim breaking out my 260018:29
mordredfungi: you do not use twitter enough18:29
clarkbmordred: neither do I. maybe this is another upside >_>18:29
mordredfungi: novel-length posts are merely broken into multiple tweets18:29
fungimordred: how do they cope with the 140-character sms message length limit these days18:30
sarobfungi: i can write a book in 140 char breaks18:30
fungioh, fragmentation. lovely18:30
harlowjareminds me of usenet files, lol18:30
sarobfungi: oh yes my friend18:30
mordredthe real fun art is making it fragments that are each self contained18:30
fungiharlowja: you should totally uuencode your messages and split them between tweets18:30
harlowjadef ;)18:30
mordredfungi, jeblair: harlowja just answered the problem18:30
mordredwe can stop using mailing lists completely18:30
clarkbsdague: looking at os loganalyze changes now. I realize I never published my draft comment from patchset 118:30
mordredand return to nntp18:30
harlowjaoh crap i can't find part 00300 (missing just that one piece of 1000), lol18:30
sdagueclarkb: hah18:31
clarkbsdague: :( it still applies to latest patchset so will publish on latest18:31
jeblairmordred: that's a really good point.  dmarc does not apply to nntp.18:31
mordredjeblair: I'm actually not even partially kidding18:31
sdagueclarkb: ok, yeh, if you do that I'll respin fast18:31
* fungi would be thrilled to set up a mail-to-news gateway18:31
jeblairmordred: old wins again.18:31
fungibeen ages since i ran a newsserver18:31
mordredmailing lists are an attempt to implement newsgroups over email that never really 100% worked well18:31
sarobjeblair: yes time updates yo18:31
sdaguenntp is actually really nice for ML18:31
mordredalso, web-gateways to nntp would be fine for folks who prefer gmail18:32
fungijust not google groups, since they dropped their nntp support a while back18:32
mordredsomething tells me we could run a web-to-nntp gateway successfully :)18:32
fungiyeah, there are plenty to choose between, from the beforetime18:33
*** lifeless1 is now known as lifeless18:33
clarkbsdague: done18:33
sdagueclarkb: oh, yep18:33
*** marcoemorais has quit IRC18:34
mordrednntp backend18:34
*** marcoemorais has joined #openstack-infra18:34
sdagueclarkb: you want to run through the rest of the stack to see if you have other stuff for me to fix?18:34
clarkbwe could just run a reddit18:34
sdagueor should I pull this part out first18:34
clarkbmordred: ^ maybe then we can convince chromakode to hack on openstack18:34
mordredclarkb: but then I'd have to learn to use reddit18:34
openstackgerritDoug Hellmann proposed a change to openstack/requirements: Update oslo.messaging to
clarkbsdague: sure let me do that really quick18:35
jeblairmordred: there's gmane18:35
sdagueclarkb: the rest of the patches I think are small18:35
jeblairfrom the guy who wrote gnus18:35
jesusaurusclarkb: yes, lets get chromakode drinking the openstack koolaid18:35
sdaguewell if we used subreddits, then openstack reactions would just be part of our normal flow18:36
mordreddoes reddit have an nntp interface?18:36
mordredjeblair: yeah - but that's the opposite way - I'd love to move to nntp and not run an email gateway for it at all18:36
*** signed8bit is now known as signed8bit_ZZZzz18:37
*** SnowDust has joined #openstack-infra18:37
SnowDustgeneral question .. who approves service gerrit accounts for third party CI ?18:37
sarobso whats a yahoo to do in the mean time?18:38
* sarob dusting off my 2600 irc cartridge18:38
SnowDustanyone ?18:39
SnowDustplz answer .. i am waiting for this frm one week now :(18:39
pleia2SnowDust: there are a few folks on the infrastructure team who do18:39
clarkbsdague: one comment on the last change in the stack18:39
clarkbsdague: but still +218:39
pleia2SnowDust: which one are you waiting on? I know SergeyLukjanov went through a bunch late last week18:39
SnowDustpleia2 i sent one mail to the team .. no one responded18:40
SnowDustits from globallogic.com18:40
* pleia2 looks18:40
SnowDustmy mail seems to be at the mailing list moderator18:40
sdagueclarkb: ok, will look18:40
pleia2SnowDust: aha, it came through on Friday, which was a holiday in the US, so please just be patient18:40
mordredthat's just crazypants18:40
pleia2SnowDust: for us it's only been a couple days :)18:40
SnowDustpleia2 : lets do it today :)18:41
pleia2the fellow who usually does them is down a laptop this week, hopefully someone else on the team will work on accounts this week (alas, I don't have those super powers)18:42
*** shayneburgess has left #openstack-infra18:43
SnowDustpleia2 : thanks for the update .. so who is the fellow behind the curtain :D18:43
SnowDustmay i know his name to bug "often" :)18:43
fungiSnowDust: bugging people doesn't provide a lot of incentive. keep in mind that those people have many, many things they're working on18:44
fungiif you annoy them, they're possibly less likely to do what you want18:45
*** marcoemorais has quit IRC18:45
*** mriedem has quit IRC18:45
*** marcoemorais has joined #openstack-infra18:46
*** marcoemorais has quit IRC18:47
*** marcoemorais has joined #openstack-infra18:47
SnowDusti want to do many many many things with that account .. but will wait18:47
*** marcoemorais has quit IRC18:47
*** marcoemorais has joined #openstack-infra18:47
*** mestery has quit IRC18:48
*** mriedem has joined #openstack-infra18:48
clarkbsdague: I am fine with that as is. lets fix the first change in the stack and hoepfully we can merge18:49
clarkbsdague: ^ is basically my comment that I left in the change18:49
sdagueclarkb: ok, I'm good with that18:50
clarkb is the pypi mirror melting down?18:51
*** mestery has joined #openstack-infra18:51
*** vhoward has left #openstack-infra18:51
*** adalbas has quit IRC18:51
sdagueclarkb: oh, hmmm... so actually.18:52
sdaguewe can't pull those regex18:52
sdaguebecause the presentation layer isn't separate yet18:52
jeblairclarkb: fungi has been looking into that (short answer so far: no)18:52
clarkbsdague: oh? can you at least import them from filter to avoid drift?18:52
*** hashar has joined #openstack-infra18:53
sdagueclarkb: yeh, or maybe I should do that presentation layer patch to complete the point of getting the structured log object18:53
sdaguelet me think about that for a few18:53
*** prad has quit IRC18:53
clarkbsdague: ok. my biggest concern is the two sets of values drifting away from each other18:54
clarkbresulint in subtle bugs that are no fun18:54
sdagueclarkb: yep, very valid concern18:54
*** prad has joined #openstack-infra18:54
fungiclarkb: so far it all seems to be from hpcloud workers, though the increase doesn't seem to strictly correspond to our increased dependence on the mirror18:54
SergeyLukjanovanteaya, I think "-1/+1 to Registered Users" isn't needed, I'll remove it in separated change, thx for catching18:54
clarkbfungi: huh18:54
fungiclarkb: so i'm starting to wonder if something is going on with network stability there18:54
clarkbjesusaurus: nibalizer ^18:54
SergeyLukjanovpleia2, SnowDust, I'm planning to go through the accounts requests tomorrow18:55
clarkbfungi: I think they may have run into similar issues18:55
*** Alexei_987 has quit IRC18:55
clarkbfungi: so wouldn't surprise me that that is the case18:55
fungiclarkb: unfortunately i'm having trouble crafting a better logstash search since it would need to be multiline to get the socket timeouts which come from pip (the socket timeouts from easy-install/setuptools i can already isolate fairly easily with a query)18:55
SnowDustSergeyLukjanov: wonderful thanks18:56
*** signed8bit_ZZZzz is now known as signed8bit18:56
fungiSergeyLukjanov: awesome. i was starting to look at it, but then realized i need to refresh my memory on the new account naming patterns18:56
*** jp_at_hp has quit IRC18:56
SnowDustfungi : thanks to u too18:57
SnowDustpleia2 : thanks thanks thanks18:57
*** SnowDust has left #openstack-infra18:57
SergeyLukjanovfungi, don't worry about it, I'll catch them tomorrow morning // presuming that now is 11 PM :)18:57
*** david-lyle has quit IRC18:58
*** david-lyle has joined #openstack-infra18:58
openstackgerritGraham Hayes proposed a change to openstack/requirements: Increased the version of oslosphinx
fungiSergeyLukjanov: yeah, have a great evening18:59
*** tusharsg has quit IRC18:59
*** david-lyle has quit IRC18:59
*** Alexei_987 has joined #openstack-infra19:00
*** praneshp has quit IRC19:00
*** david-lyle has joined #openstack-infra19:01
*** praneshp has joined #openstack-infra19:03
*** dims_ has quit IRC19:03
*** praneshp has quit IRC19:03
*** praneshp has joined #openstack-infra19:04
nibalizerfungi: so what are you seeing re: network instability in hp cloud? I'm stepping into this conversation midway and having some trouble catching up19:04
*** e0ne has quit IRC19:05
funginibalizer: socket timeout errors from both pip and easy-install/setuptools when retrieving packages from our mirror. it could be a problem with hpcloud's network, or a problem at rackspace's edge, or it could just be a coincidence that we're not hitting it with workers in rax regions19:05
funginibalizer: http (80/tcp) specifically19:06
fungisince we haven't switched it to https yet19:06
nibalizerare you seing connection reset by peer in the logs for when jenkins tries to run pip install?19:06
*** adalbas has joined #openstack-infra19:07
funginibalizer: occasionally, though not as often (i've not run numbers on that failure mode to compare, just going on gut)19:07
nibalizercould you provide me to a link to the console output for when this dies?19:07
fungisure, there are a couple different ways depending on whether it dies on a requirement or a setup_requires19:08
* fungi digs up examples19:08
nibalizeri'd love to see all of them :)19:08
funginibalizer: here's one where it hit while easy-install/setuptools was trying to pull down a setup_requires
nibalizeroh thats a legit socket timeout19:10
nibalizernot someone hanging up the phone19:10
fungisomeone cut the line ;)19:11
* jesusaurus reads backscroll19:12
KiallSo - What are chances of convincing you guys to make WIP stick across patchsets? :)19:13
funginibalizer: here's one which struck pip directly
jeblairKiall: slim -- it breaks the ability for core reviewers to set wip19:15
*** e0ne has joined #openstack-infra19:15
fungior more to the point, breaks the ability for change owners to un-set wip19:15
jeblairfungi: yes19:15
*** shashankhegde has quit IRC19:15
*** sweston_irccloud has joined #openstack-infra19:16
KiallLetting core reviewers WIP other peoples code breaks the ability to have WIP stick ;) Fair enough, I see the reason why it doesn't now..19:16
fungiat least with the current label vote based implementation of wip, where it's not a change status19:17
*** e0ne has quit IRC19:18
jesusaurusfungi: those are indeed strange timeouts, i havent seen anything like that. you only see them from hpcloud nodes?19:18
sdagueKiall: the nova hack for that for core reviewers is to -2 your own patch set19:18
*** andreykurilin_ has joined #openstack-infra19:18
Kiallsdague: Yea - that would work actually, as CR-2 sticks19:19
*** arnaud__ has quit IRC19:19
clarkbI have restarted the decoupled logstash indexers on worker19 and 2019:19
clarkbif my hunch is correct we will see them pick up the pace post restart. And if that happens I suppose I will have to look at upgrading logstas19:19
*** e0ne has joined #openstack-infra19:19
*** e0ne has quit IRC19:20
fungijesusaurus: specifically, we only seem them from hpcloud nodes trying to reach our pypi mirror in rackspace (and not from rackspace nodes in other regions than the pypi server)19:20
clarkbis there anyway to query the neutron routers for info on packet loss and the like?19:21
fungijesusaurus: at least in the past couple days i've been looking at, so if we are seeing them from workers in rackspace it's far, far fewer than the proportion of rackspace to hpcloud workers would suggest if it were more general19:21
clarkbhow much info is exposed by the neutron router?19:21
clarkbmestery: ^19:22
clarkbif 19 and 20 do not show improvements tat indicates the elasticsearch client daemon is the process struggling and I will restart that process on 17 and 1819:23
funginibalizer: i've also spotted some "network unreachable" (not sure if this is a happy eyeballs sort of fallback to ipv6 even though there's no route/gateway for it or what)
* mestery reads scrollback19:24
fungiDownload error on [Errno 101] Network is unreachable -- Some packages may not be found!19:25
clarkbmestery: long story short there is network weirdness in hpcloud which runs neutron and was wondering what sorts of diagnostic tools are available to the customer19:25
mesteryclarkb: Understood, checking for you now, apologies, in the mid-cycle at the same time right now :)19:26
clarkbmestery: no rush, we can probably dig around ourselves but I figured going to the source might be quicker :)19:26
mesteryclarkb: ;)19:26
mesteryclarkb: I don't think there's a way to do that, in fact, I found a BP which was registered proposing this but not implemented:
* mestery takes a note to look into this.19:28
funginibalizer: yep, these were all hpcloud too, though admittedly so few that i don't feel there's a representative sample...19:28
clarkbmestery: thanks19:28
clarkbmestery: ya router-show seems pretty basic19:28
openstackgerritCraig Bryant proposed a change to openstack-infra/config: Add Monasca sub projects
nibalizerfungi: okay well thanks a ton for grabbing the errors19:28
funginibalizer: yw19:29
clarkbsame with net shows19:29
clarkbfungi: we can relatively easily set up some experiments19:29
fungiclarkb: indeed we can. we are scientists19:29
mesteryclarkb: Yup, sorry. :(19:30
clarkbfungi: boot a couple nodes on differen routers that we already have setup. Then pipe /dev/urandom on each box to the other via nc19:30
clarkbfungi: or similar and see what we end up with19:30
jeblairhere is a POC of the table-of-test-results idea in gerrit, implemented on top of the hideci javascript hack:
clarkbif that is all happy we will then need to add an external node to do that to19:30
jeblairclarkb, fungi, mordred, sdague: ^19:30
clarkbjeblair: NICE19:30
jeblairneeds some styling, but that's the bulk of the functionality19:31
*** marcoemorais has quit IRC19:31
*** marcoemorais has joined #openstack-infra19:31
jeblairalso, probably some adaptation to also work in the new change screen19:31
sdaguejeblair: hmmm... I don't see anything. I'm getting a jquery error19:31
fungiclarkb: i feel like we'd probably need to establish a bunch of http connections and turn them over regularly to simulate the circumstances19:32
*** jlibosva has joined #openstack-infra19:32
fungiclarkb: similarly between hpcloud and rax and within rax and compare numbers19:32
jeblairsdague: try shift-reloading, or re-navigating to it.  i think there may be a problem with the hideci inclusion of jquery19:32
clarkbfungi: well I want to isolate the issue as close to hpcloud if possible hence nodes on different routers19:32
*** blamar has joined #openstack-infra19:32
clarkbfungi: that will not traverse the nat19:32
clarkband the external networking19:32
*** asettle has quit IRC19:32
jeblairsdague: (so we may want to test that a bit before we merge hideci)19:33
*** asettle has joined #openstack-infra19:33
clarkbsdague: ya I had to hard reload19:33
fungiclarkb: yeah, that would help us rule out neutron, but i still have a hunch there's a "web accelerator" or something at hpcloud's border which could be to blame19:33
sdaguehmmm... I've shift reloaded a couple of times19:33
*** marcoemorais has quit IRC19:33
fungiclarkb: or at rackspace's border for that matter19:33
sdague"Uncaught TypeError: Cannot read property 'length' of undefined "19:33
jeblairsdague: oh, that may be a js compat bug; i only tested ff19:34
sdagueyeh, I brought up ff, and same issue there19:34
*** marcoemorais has joined #openstack-infra19:34
sdaguecan someone screen shot?19:34
sdaguebecause it looks like a normal change to me at the moment :)19:35
openstackgerritJoshua Harlow proposed a change to openstack-infra/config: Add templated uploading jobs for doc8 project
fungisdague: do you not see the "toggle ci" button at the bottom?19:35
fungisdague: i think maybe it also didn't work if i wasn't logged into review-dev, but not positive19:35
sdagueI logged in19:35
clarkbsdague: ctrl + shift + r? works for me in ff19:35
sdagueI don't see a button19:35
fungiyeah, when i shift+r in firefox i get it whether i'm logged in or not, it seems19:36
*** sarob has quit IRC19:36
nibalizerjeblair: looking sweet!19:37
*** mpavone has quit IRC19:37
*** jlibosva has quit IRC19:37
*** miqui_ is now known as miqui19:37
fungijeblair: yep, neat-o19:37
*** Sukhdev has quit IRC19:38
sdagueoh, ok, got it in firefox now19:38
*** amitpp has joined #openstack-infra19:39
sdaguethough the ci comments are still inline19:39
fungineeds a "hide puny humans" button to go along with it19:40
*** shashankhegde has joined #openstack-infra19:40
jeblairsdague: yeah, i think if we do the table, we'd probably want it to default to hiding the ci comments19:40
*** mbacchi has quit IRC19:41
sdaguejeblair: so where I'd personally really like the table is to the right of the humans. So basically you have 2 columns, humans on left, robots on right.19:41
reedfungi, what's the status of hiera keys for ?19:42
jeblairsdague: agreed19:42
jeblairsdague: just seemed like a good stopping point before fiddling more with css19:42
*** mrmartin has joined #openstack-infra19:42
fungireed: funny you should ask--just got back to adding them on the puppet master now19:43
fungireed: eta a few minutes19:43
reedmrmartin, check ^^19:43
reedfungi, thanks, great19:43
mrmartinfungi: hi19:44
jeblairsdague: the roulette of the jquery include is disturbing though; i'm not sure what to do about that (i just tried moving it to the footer, but no change)19:44
fungireed: mrmartin: and then i can try bringing up the groups.o.o server with that change19:44
*** tsg has joined #openstack-infra19:44
sdaguejeblair: yeh, I don't know19:44
sdaguejeblair: is it interfering with jquery in gerrit?19:45
jeblairsdague: gerrit doesn't use jquery (last i checked)19:45
sdaguewell jquery min might be stomping on autominified vars for gerrit19:45
*** homeless has quit IRC19:46
*** mestery has quit IRC19:46
sdaguethere is a way to instantiate jquery not into the global namespace19:46
*** e0ne has joined #openstack-infra19:46
*** dims has joined #openstack-infra19:46
*** e0ne has quit IRC19:46
*** mestery has joined #openstack-infra19:47
*** homeless has joined #openstack-infra19:48
*** sarob has joined #openstack-infra19:49
*** mestery has quit IRC19:49
arosenHi, Sorry to bug you guys you guys but it would be awesome if someone could check out this conf change when someone gets a chance :)
*** masayukig has quit IRC19:50
anteayaSergeyLukjanov: actually according to fungi the +1/-1 to Registered Users on stable/icehouse for sahara is needed, since you are overwritting the default to stable-maintenance team19:50
*** sarob_ has joined #openstack-infra19:50
jeblairsdague: oh, that may be my error19:51
SergeyLukjanovanteaya, oh, okay, /me should remember it19:51
fungianteaya: SergeyLukjanov: well, it's needed if you are setting exclusive group permission for code review19:51
anteayaSnowDust, actually let's start by finding out what it is you want to do with the account "many things" has a ring to it that sits uneasily with me19:52
SergeyLukjanovfungi, yeah, got it19:52
*** amitpp has quit IRC19:52
fungianteaya: SergeyLukjanov: which you'd have to do to override the blanket one we have for refs/stable/* in all-projects granting -2/+2 to openstack-stable-maint19:52
*** mkoderer has quit IRC19:52
anteayaSergeyLukjanov: yes let's find out what this person wants to do with their third party ci system19:53
*** sarob has quit IRC19:53
anteayaSergeyLukjanov: I have interacted with this person before and my gut response is they don't show good judgement in their decisions19:53
anteayaSergeyLukjanov: so now this person wants a presence in the third party space, I would like more details19:54
*** dims_ has joined #openstack-infra19:55
SergeyLukjanovanteaya, yup, more details is good19:55
*** masayukig has joined #openstack-infra19:55
*** Sukhdev has joined #openstack-infra19:55
fungianteaya: if you're talking about SnowDust, the subject on the e-mail suggested it was for testing changes to trove i think19:56
fungijust didn't re-mention that in the message body19:56
*** jcoufal has quit IRC19:56
jeblairsdague: yep, found it.  my bug.  should be better now.19:56
*** dstanek_zzz is now known as dstanek19:56
*** dims has quit IRC19:57
*** morganfainberg_Z is now known as morganfainberg19:57
sdaguejeblair: ok, so after that I get the toggle ci button, but not the table in chrome, fwiw19:57
jeblairsdague: yeah, in chrome i find i need to navigate away and back; probably something with the onload function19:58
openstackgerritJames E. Blair proposed a change to openstack-infra/config: WIP: add CI results table
*** prad has quit IRC19:59
*** arnaud has joined #openstack-infra20:02
jeblairthere's what i have so far; /me lunches20:02
*** hashar is now known as hasharMeeting20:02
*** thedodd has quit IRC20:03
*** thedodd has joined #openstack-infra20:05
*** enikanorov__ has joined #openstack-infra20:05
*** marcoemorais has quit IRC20:06
*** YorikSar has quit IRC20:06
*** enikanorov has quit IRC20:06
*** marcoemorais has joined #openstack-infra20:06
*** marcoemorais has quit IRC20:07
openstackgerritSergey Skripnick proposed a change to openstack-infra/config: Fix file name and path of rally html results
*** marcoemorais has joined #openstack-infra20:08
*** YorikSar has joined #openstack-infra20:08
anteayafungi: thanks yes, was speaking about SnowDust, I'm looking for his email now20:10
*** afazekas has quit IRC20:10
openstackgerritCraig Bryant proposed a change to openstack-infra/config: Add Monasca sub projects
openstackgerritJeremy Stanley proposed a change to openstack-infra/config: Add instance
anteayafungi: found it, going to chat with SlickNik20:12
*** jergerber has joined #openstack-infra20:14
*** skraynev has quit IRC20:14
*** krtaylor is now known as krtaylor_away20:14
*** prad has joined #openstack-infra20:14
arosenHi, from time to time  I run into this situation with testr where it outputs the results like this: .  I was wonder the best way to debug this? I'm guessing there is a syntax error somewhere in the source tree which is causing this?20:17
*** nati_uen_ has joined #openstack-infra20:18
fungiarosen: it's an import error, fixed in (i think) testr upstream but no new release yet20:18
fungiwell, the formatting of import errors in the output stream is fixed upstream20:18
fungiyour actual import error is, of course, still an error20:18
clarkbfungi: correct20:19
*** jamielennox is now known as jamielennox|away20:19
clarkbarosen: if you look towards the end of the nasty string there should be something like import errors and a list of files that failed to import20:19
*** skraynev has joined #openstack-infra20:19
fungiarosen: looks like congress.tests.test_server is the culprit20:19
*** mestery has joined #openstack-infra20:20
*** signed8bit has quit IRC20:20
*** nati_ueno has quit IRC20:20
ryanpetrelloany chance somebody could take a look at ?20:21
ryanpetrelloI have a pecan bug fix I need to get released, and I’m blocked by some failing tests in stable ironic20:21
arosenit doesn't say what it failed to import though right? My pyflake checker isnt' seeing anything wrong20:21
*** rcarrillocruz has joined #openstack-infra20:22
arosenlet me try to import what i import by hand.20:22
*** radez is now known as radez_g0n320:22
rcarrillocruzhey folks, i'm importing a git repository with a lot of branches to a gerrit project20:22
clarkbarosen: right I think lifeless said the python discover protocol doesn't expose that in a clean way :(20:22
rcarrillocruzadded myself to a group with create reference, forge commiter, forge author20:22
clarkbryanpetrello: dne20:22
rcarrillocruzdoing the push like this:20:22
clarkbrcarrillocruz: er done20:22
arosenhrm i can import everything that's imported in this file :/20:22
*** markmcclain has quit IRC20:23
fungiarosen: in the tox venv?20:23
rcarrillocruzgit push -f <remote> *:*20:23
ryanpetrelloclarkb, fungi: thanks!20:23
rcarrillocruzi see it imported in the gerrit master20:23
rcarrillocruzand all tags20:23
rcarrillocruzbut not other branches20:23
fungiarosen: or do you maybe have something installed locally which it needs and is missing in the venv?20:23
*** signed8bit has joined #openstack-infra20:23
clarkbrcarrillocruz: is there some reason you didn't use jeepyb's manage-projects to do it?20:24
sdaguefungi: looks like another instance of the pip download issue20:24
*** hasharMeeting is now known as hashar20:24
sdaguethough different message20:24
clarkbrcarrillocruz: but I think you may need one of the special characters in your refs to make it create new refs?20:24
openstackgerritA change was merged to openstack-infra/config: Add instance
clarkbsdague: its looking like logstash is the problem20:24
sdagueclarkb: oh....20:24
clarkbsdague: restarting just the logstash bit restores what I expect as normal throughput for the worker20:24
sdagueclarkb: what is logstash doing?20:25
clarkbsdague: over time it just bogs down and stops moving data20:25
rcarrillocruznot doing it in upstream, here at hp... first time i do, i don't know if i'm running the correct git push options20:25
sdagueclarkb: interesting20:25
clarkbrcarrillocruz: I understand, is there some reason you didn't use manage-projects?20:25
*** e0ne has joined #openstack-infra20:25
arosenfungi:  there should be a .venv dir which is created as the result of running tox -epy27 ?20:25
clarkbrcarrillocruz: basically we made that tool so that we don't have to think about it20:25
sdaguearosen: .tox/py2720:25
fungircarrillocruz: that's how m-p gets the heads duplicated, but you also need to duplicate tags and HEAD with separate steps20:25
*** e0ne has quit IRC20:25
fungiarosen: in .tox/envname20:26
clarkbsdague: it may be related to garbage collection? maybe it is leaking and derping somewhere20:26
rcarrillocruzok, so not so simple as gerrit documentation states20:26
clarkbsdague: anyways the good news I can look at upgrading logstash20:26
rcarrillocruzwill look at manage_projects , thx20:26
*** mestery has quit IRC20:26
sdagueclarkb: cool20:26
clarkbsdague: because I have decoupled it from our version of elasticsearch20:26
*** sweston_ has quit IRC20:27
*** Ajaeger1 has quit IRC20:28
*** shayneburgess has joined #openstack-infra20:28
*** shayneburgess has quit IRC20:28
arosenfungi:  should there be an activate file there or a venv file ?20:28
*** e0ne has joined #openstack-infra20:29
clarkbarosen: thats the root of the virtualenv env. so the activate file is at .tox/envname/bin/activate20:29
fungiarosen: or you can just interactively invoke .tox/envname/bin/python20:30
arosenclarkb:  all i have in .tox is: dist  log  pep8  py2720:30
fungiarosen: yeah, so .tox/py27/bin/python20:30
fungior . .tox/py27/bin/activate20:30
fungior however you prefer to work with your virtualenvs20:30
arosencool got it thanks20:30
ryanpetrelloclarkb: is this failure just a fluke?20:31
ryanpetrelloor something I’ve done wrong?20:31
clarkbryanpetrello: I don't think its related to what you did20:31
fungiGot 1 failure(s) while initializing: change from absent to directory failed: Could not set 'directory on ensure: File exists - /var/run/puppet20:31
arosenfungi: clarkb awesome that shows me it fails there :)20:31
clarkbryanpetrello: your change passed the same test at check time20:31
clarkbunless we got something that slipped through?20:32
fungiarosen: yep, probably missing a requires or test-requires which you have installed on your system20:32
fungiclarkb: looks to me like maybe the puppet-apply test could have a race condition somewhere20:33
clarkbfungi: ohhhhh good point20:33
arosenfungi:  yup that's exactly it. I added at test for eventlet which imported eventlet but i didn't have that in requirements.txt20:33
clarkbfungi: ya I bet that is what it is20:33
fungiclarkb: maybe it needs some debug output20:33
clarkbfungi: well last time mordred tried that it didn't help :)20:33
arosenfungi:  clarkb thanks!20:33
clarkbfungi: I think we may need to have puppet runs use their own /var/run dirs?20:33
fungiclarkb: ooh, good point. could be contention over it?20:34
clarkbfungi: ya thats what I am thinking20:34
fungiryanpetrello: looks like you found a bug in our test methodology20:34
* fungi opens a bug report about this20:35
*** pballand has quit IRC20:35
mordredwhat did I do?20:36
jeblairfungi, clarkb: so why is there _sometimes_ a /var/run/puppet?20:36
clarkbmordred: you turned on debug puppet output on that test and it didn't help you :)20:36
mordredclarkb: that is correct20:36
mordredfungi: puppet apply debug output is a pile of useless20:37
clarkbjeblair: xargs is running the processes in parallel. I am guessing that the contention for that dir is low?20:37
jeblairclarkb: gotcha20:37
clarkbbut I don't know a lot about puppets state model20:37
fungiryanpetrello: reverify bug 133987520:37
uvirtbotLaunchpad bug 1339875 in openstack-ci "Parallel puppet-apply jobs fail creating existing /var/run/puppet" [High,Confirmed]
jeblairclarkb: why is puppet --noop creating dirs?20:37
ryanpetrellofungi: thx20:38
clarkbjeblair: I think its creating state dirs that puppet needs to run20:38
*** YorikSar has quit IRC20:38
clarkbjeblair: its not creating them as a response to file{} resources20:38
mordredand it's just using its own engine to accomplish that20:38
*** prad has quit IRC20:38
jeblairclarkb: it just _looks_ like it's doing that.  :)  (as mordred said)20:38
fungiclearly puppet's definition of "no operation" and mine differ slightly20:38
mordredjeblair: well, there is NO WAY to know what it's actually doing20:38
jeblairmordred: i hate quantum computing20:39
openstackgerritA change was merged to openstack-infra/config: Mark OpenStack stable branches as non-voting for pecan; they're too unreliable.
* fungi collapses the puppet wave function20:39
*** YorikSar has joined #openstack-infra20:40
*** e0ne has quit IRC20:42
*** andreykurilin_ has quit IRC20:42
mordredthat's the funniest commit message20:42
clarkbmordred: I think ryanpetrello found the bits had all rotted20:43
*** bknudson has joined #openstack-infra20:43
openstackgerritpatrick-crews proposed a change to openstack-infra/elastic-recheck: Initial work to add query namespace to elastic-recheck.
ryanpetrelloyea, unfortunately :\20:43
ryanpetrelloI’m glad to contribute to OpenStack projects, it’s part of my job :)20:44
jeblairthe stable jobs haven't been run in 2.5 months and no one noticed20:44
ryanpetrellobut I can’t have pecan help up waiting on backports that sometimes take weeks :|20:44
mordredjeblair: but stable branches are important!!!!!20:44
mordredjeblair: people CARE!!!!!20:44
ryanpetrelloyea, I spoke w/ devananda20:44
jeblairmordred: in fact, we're still not running them
ryanpetrelloand the Ironic folks, at least, seemed to express interest in having some sort of nightly (or regularly) schedule run of stable branches20:45
ryanpetrellojust so that e.g., if a client breaks something, you know that stable’s breaking, too20:45
openstackgerritpatrick-crews proposed a change to openstack-infra/elastic-recheck: Initial work to add query namespace to elastic-recheck.
sdaguejeblair: only the stable periodic jobs20:45
* clarkb approves20:45
openstackgerritCraig Bryant proposed a change to openstack-infra/config: Add Monasca sub projects
*** arnaud has quit IRC20:45
*** sarob_ is now known as sarob20:45
fungireed: mrmartin: there's an existing server, i forgot, so the puppet update should be in place on it now. do you need me to initialize the database or anything, or should it take care of that automatically?20:45
ryanpetrellountil such a thing exists, though, I can’t keep playing wac-a-mole :\20:45
sdaguewe're running icehouse on every tempest change now with branchless20:45
jeblairryanpetrello: yeah, we totally have those jobs.  someone broke them 2.5 months ago and no one noticed.20:45
*** prad has joined #openstack-infra20:45
*** andreykurilin_ has joined #openstack-infra20:45
ryanpetrellojeblair: ha, didn’t catch that20:45
mrmartinfungi: we need just an empty db20:46
fungimrmartin: okay, that should be hooked up for you then20:46
clarkbjeblair: after the last summit I had hope again that someone would care :/20:46
*** e0ne has joined #openstack-infra20:46
clarkbjeblair: oh well20:46
*** maxbit has quit IRC20:46
mrmartinfungi: thanks20:46
clarkbanyways change is approved hopefully will be better soon20:46
*** maxbit has joined #openstack-infra20:46
sdagueclarkb: yeh, speaking of which, can we bring up this crazy pants idea of icehouse having extended support on the list?20:46
fungimrmartin: looks like this was the update getting applied:
clarkbsdague: extended support?20:47
*** mfer has quit IRC20:47
rcarrillocruzclarkb, fungi: i didn't have all branches checked out locally, no wonder they weren't pushed20:47
clarkbhonestly if someone wants to extend support on something lets do it when python26 dies20:47
clarkbeg kilo20:47
sdagueI thought someone told me that in the stable branch session they wanted an extra 6 months on icehouse20:47
clarkbsdague: oh right. I think we told people to step up20:47
clarkbbut that hasn't happened so meh20:48
sdagueyeh, as long as our policy remains, if you don't show up, we don't care, I'm good with it :)20:48
fungiclarkb: line 50 of totally means people will magically step up and keep it running, right?20:48
clarkbright so 15 months is a 3 month contraction right?20:49
clarkbwe did 18 months previously20:49
*** salv-orlando has joined #openstack-infra20:49
fungiwe did 9 on grizzly20:49
fungiand planned to do 9 on havana as well20:49
fungibut pushed icehouse to 1820:49
sdagueyeh, it's a 6 month extension20:49
fungier, 1520:49
clarkbfungi: we did 1520:49
clarkb9 + 6 dev20:49
clarkb= 1520:49
fungiclarkb: oh, this is 15 from release20:50
clarkbthats how I count20:50
clarkber what?20:50
fungiheh, then 21 by your count20:50
clarkbno I wasw in that room20:50
mordrednibalizer, AaronGr: what are the downsides, if any, to having puppetdb api ports open?20:50
clarkbeither I completely misunderstood20:50
sdagueclarkb: so it's good that it's completely clear :)20:50
clarkbor we had people talking past each other20:50
clarkbbecause when I talk about support I talk about entire lifetime of a thing20:50
clarkbthat includes its time on master20:50
fungiyeah, i assumed we were counting from release day, not from start of development cycle20:51
clarkbbecause that determines how long I need to keep a test bed running for it20:51
*** maxbit has quit IRC20:51
jeblairi'm with fungi on this; we don't "support" something that isn't released yet20:51
fungiin this case "stable support for 15 months" implies it's stable and released for 15 months20:51
*** maxbit has joined #openstack-infra20:51
clarkbI clearly misunderstood then20:51
fungiinfra may support the test infrastructure for those things, but the stable branch managers and vmt do not20:52
clarkbbut I think it is still reasonable to say its month 12 and its broken no one is fixing it. delete branches20:52
fungiyeah, we said it would be contingent on actual resources20:52
sdaguemaybe we should get our fearless release manager to clarify20:52
*** e0ne has quit IRC20:52
fungihopefully he's asleep now20:53
clarkbfungi: jeblair sdague I think I am happy with now20:53
sdaguefungi: so... I'm very skeptical of people showing up for this. Because if they aren't engaged way ahead of the game, someone showing up at the 11th hour for stable only is not useful.20:53
clarkbonce we have that my next step will be to work on upgrading the logstash indexers on the workers20:53
fungisdague: totally agree20:53
clarkbfungi: jeblair sdague the reason that helps is we don't have to upgrade elasticsearch in lockstep if the logstash indexer is speaking http to a local elasticsearch daemon that joins the cluster20:54
clarkbonly the thing that joins the cluster must be upgraded in lockstep20:54
sdagueclarkb: nice20:54
fungiclarkb: lgtm20:54
fungii'm heading out to dinner. inlaws have come to see the new place so will probably be afk a while20:55
clarkbfungi: have fun20:55
*** reed has quit IRC20:55
clarkbyou can see that when I restarted the indexer around 19:20 that things picked up at
*** juzuluag has quit IRC20:56
wenlockhave you guys run into this error this week?  with jenkins job builder20:56
wenlocklatest git changes20:56
*** juzuluag has joined #openstack-infra20:56
*** reed has joined #openstack-infra20:56
wenlocki don't see that the six library is in the requirements.txt either20:56
clarkbwenlock: looks like you didn't paste everything20:56
wenlockclarkb this one
clarkbwenlock: the import of six is happenign in a different module not in JJB20:58
*** maxbit has quit IRC20:58
*** sarob has quit IRC20:58
clarkbwenlock: so its fine that JJB doesn't include it. How did you install? I would expect the deps to install their deps too20:58
*** maxbit has joined #openstack-infra20:58
sdagueclarkb: you know, as much as I like puppet, there are times when I want to punch them for the conflicting resource definition issue20:58
openstackgerritA change was merged to openstack-infra/config: Fix stable branch selector in zuul layout
clarkbwenlock: is the problem20:58
clarkbwenlock: six is listed as a test requirement but it should be a requirement20:59
clarkbzaro: ^20:59
clarkbsdague: especailly when they are the same20:59
wenlockclarkb using vcsrepo for then pip install /opt/jenkins_job_builder exec20:59
*** sweston__ has quit IRC20:59
clarkbsdague: would be nice if you could just merge the two20:59
clarkbwenlock: ya I think the bug is in python-jenkins20:59
clarkband it wasn't caught because the tests worked20:59
* clarkb will push a change21:00
wenlockclarkb awesome21:00
*** e0ne has joined #openstack-infra21:00
*** juzuluag has quit IRC21:01
wenlockclarkb if i want a local quick test, is it just latest version of six that will do the trick?21:01
*** dkliban is now known as dkliban_brb21:01
*** thedodd has quit IRC21:01
*** e0ne has quit IRC21:01
clarkbwenlock: yes21:01
wenlockclarkb ty21:01
wenlockclarkb yeahy :D JJB is happy again21:02
wenlockclarkb let me know if you need me to test it again on our side once you patch21:02
*** jamespd_ is now known as jamespd21:02
openstackgerritClark Boylan proposed a change to stackforge/python-jenkins: Six is a runtime requirement not a test req
clarkbzaro: jeblair wenlock ^ that should do it21:04
hasharclarkb: :-(21:04
mordredsdague: the conflicting resource thing is the most braindead thing ever21:05
*** prad has quit IRC21:06
wenlockclarkb thanks!21:07
*** dustins has quit IRC21:07
clarkbhashar: you are awake! watching argentina and the netherlands?21:08
hasharI don't watch baseball21:08
*** obondarev_ has joined #openstack-infra21:09
clarkbI forgot to git add the requirements file :/21:09
*** dprince has quit IRC21:09
openstackgerritClark Boylan proposed a change to stackforge/python-jenkins: Six is a runtime requirement not a test req
hasharclarkb: reference the commit message that introduced the error while at it : 5d7ecd4321:09
clarkbran tests locally even but didn't do a fresh fetch21:10
hasharwas about to nitpick about adding a reference21:10
*** prad has joined #openstack-infra21:10
clarkbhashar: I don't think that is necessary21:10
hashar(wikipedia biased)21:10
clarkbhashar: you can always look in git log21:10
hasharI am lazy ;-D  But fine to me21:10
hasharrelevant xkcd :
hasharwe actually had a beer brewed just for us that had that name21:11
*** maxbit has quit IRC21:11
clarkbI want to brew more beer21:11
*** maxbit has joined #openstack-infra21:11
*** obondarev__ has joined #openstack-infra21:11
hasharI want to start brewing myself21:11
clarkbhashar: I like it. teaches me patience21:11
*** jamielennox|away is now known as jamielennox21:11
hasharwife object because it stinks (she must know about FUD)21:11
clarkbits actually not that bad21:11
hasharkids++ is a great wait to exercise your patience21:12
clarkbespecialyl if you put the fermentor in a closet or out of the way. it won't stink up a whole house21:12
*** julim has quit IRC21:12
*** obondarev has quit IRC21:12
*** obondarev_ has quit IRC21:12
*** obondarev__ is now known as obondarev21:12
pleia2only really stinks if you're doing all grain anyway21:13
pleia2more of concern is other stinks around, had the strangest brew on one day when the lawn was mowed the same day as brewing (pro tip: don't do that)21:13
clarkbpleia2: I find all grain doesn't make it stinkier21:14
pleia2maybe it just lasts longer :)21:14
clarkbyou do make things smell like bread during the mash though :)21:14
hasharclarkb: do we care about six minimal version?  python-jenkins uses six.moves.urllib which was introduced in six 1.4.021:15
Mithrandiruhm, brewing shouldn't stink21:15
clarkbhashar: I don't think so. I copied what was already there21:15
Mithrandirunless you're doing something very weird.21:15
hasharclarkb: yeah I am just paranoid for now reaon21:15
clarkbMithrandir: I think people think that way because sulphury stuff stinks21:15
clarkbMithrandir: ciders and wines iirc21:15
*** mmaglana has quit IRC21:16
Mithrandirclarkb: there's no to very little H2SO4 in beer unless you've fucked up.21:16
*** mmaglana has joined #openstack-infra21:16
Mithrandirin which case, stop fucking up. :-P21:16
hasharI will request a new documentation project openstack-infra/brewing to get nice receipes at
clarkboh I agree, just saying I think that is where people get that idea21:17
pleia2hashar: +121:17
clarkbI have heard cider get described as "rhino farts"21:17
clarkbbeer isn't like that21:17
pleia2Mithrandir: some people really don't like the smell of the mash21:17
*** timrc is now known as timrc-afk21:17
clarkbpleia2: wut21:17
pleia2maybe they don't like beer21:17
clarkbblashpemy mashing is one of the best parts21:17
hasharsleeping time. see you later!21:17
Mithrandirpleia2: some people suck. :-/21:18
*** hashar has quit IRC21:18
*** maxbit has quit IRC21:19
*** tsg has quit IRC21:19
*** maxbit has joined #openstack-infra21:19
*** sarob has joined #openstack-infra21:19
clarkbbut ya I have never had problems. in the fall, winter, spring (and summer when brewing saisons) I use a closet that has relatively stable temps. Other summer beers go in the keezer to ferment. Biggest issue is when you don't use a blow off when you should and have to clean up yeast off of everything21:20
clarkbone time a lid exploded off the bucket that was fun21:20
*** timrc-afk is now known as timrc21:20
asselindoes zuul create this ref within gerrit? if so what permissions would it need? 2014-07-09 14:05:18,930 DEBUG zuul.Repo: CreateZuulRef master/Z8f5de7f72c344317a07cffa1c0e83317 at 5e9b9682e43dfe8fc608eaac4a9def1182e986ca21:21
clarkbasselin: it did in the past but now it doesn't by default (I beleive it is an option to push that back though)21:22
*** markmcclain has joined #openstack-infra21:22
clarkbasselin: instead it should be served by the zuul merger21:22
*** shayneburgess has joined #openstack-infra21:22
asselinok I you should get access to it via git fetch -t ssh://zuul-merger-ip...?21:23
clarkband last thing before I stop talking about brewing :)21:23
*** mrmartin has quit IRC21:23
clarkbgigantic has released all of their recipes which is super cool21:23
clarkbasselin: yup21:23
clarkbasselin: we use http instead though21:23
clarkbbut ssh or git:// would work depending on your setup21:24
ryanpetrelloclarkb: draughtcraft.com21:24
ryanpetrellopersonal webapp of mine (running on pecan and DH’s public cloud)21:25
openstackgerritA change was merged to stackforge/python-jenkins: Six is a runtime requirement not a test req
clarkbryanpetrello: neat21:25
pleia2ryanpetrello: hah, neat21:25
clarkbryanpetrello: I have been using beersmith but will have to give that a go21:25
ryanpetrelloyea, beersmith is definitely more full-featured/scientific21:25
ryanpetrelloI got tired of all the knobs and built something pretty simple21:25
ryanpetrelloplus, that’s OpenStack under the hood :D!21:26
*** thedodd has joined #openstack-infra21:26
clarkbryanpetrello: thats really cool. The knobs that I really like from beersmith are the mash knobs21:26
ryanpetrelloyea, none of that in my software21:26
ryanpetrelloI just have a spreadsheet I use for mashing21:26
clarkbryanpetrello: though I basically have it dialed into where I have some rough imperial units for specific mash temps21:26
ryanpetrellosame, I can mostly scratch it out on paper at this point21:27
clarkbon my system its something like mashing at 152F is 17quarts of 166F water into about 9-13lbs of grain21:27
openstackgerritmelanie witt proposed a change to openstack-infra/elastic-recheck: Extend query for pip download timeout bug 1270710
uvirtbotLaunchpad bug 1270710 in openstack-ci "sporadic pip timeouts during download" [Medium,Expired]
clarkbthen sparge with 19quarts iirc21:27
*** aysyd has quit IRC21:28
*** dimtruck is now known as zz_dimtruck21:29
ryanpetrelloclarkb, pleia2: best part about my app:
clarkbooh open source21:29
pleia2yay :)21:29
*** shivharis has joined #openstack-infra21:30
jeblairryanpetrello: does it have a license?21:30
*** maxbit has quit IRC21:30
*** praneshp has quit IRC21:31
*** maxbit has joined #openstack-infra21:31
openstackgerritCraig Bryant proposed a change to openstack-infra/config: Add Monasca sub projects
ryanpetrellojeblair: it will in a moment, good catch21:31
clarkb I have a beer very similar to that sitting in a fermenter right now21:32
*** tsg has joined #openstack-infra21:32
openstackgerritK Jonathan Harker proposed a change to openstack-infra/config: Update git refs before comparing refs
*** zehicle_at_dell has joined #openstack-infra21:32
clarkbbut I used more hops >_>21:33
shivharisclarkb: how can i retrigger third party ci, if i realize that a trigger was missed - due to error/testing of a CI testbed21:33
clarkbshivharis: it depends on the third party ci. I think most of them honor our recheck syntax21:34
ryanpetrelloclarkb: yaaaay smash21:34
shivharisclarkb: i tried "recheck no bug" none got restarted and my own did not get a trigger21:34
shivharisclarkb: am i using the correct syntax for recheck21:35
clarkbryanpetrello: ya I really like simple beers. I went through a period where I tried doing too much with beer and there were always issues. One of the best beers I have ever brewed was a "leftovers" saison batch 11lbs pilsner malt, 2 oz saaz, wyeast 3711 so good in the summer21:35
clarkbshivharis: that is the correct syntax for upstream tests21:36
clarkbshivharis: but the third party tests may not honor it21:36
*** gondoi is now known as zz_gondoi21:36
shivharisclarkb: thanks for confirming my syntax21:36
clarkbryanpetrello: and the printing works!21:38
clarkbryanpetrello: ok this is nice21:38
ryanpetrellouse it, it’s free :D21:38
*** zehicle_at_dell has quit IRC21:38
ryanpetrelloand if there’s something wrong, pull request21:38
ryanpetrelloonly downside, it doesn’t import currently21:38
ryanpetrellobut it does export21:38
clarkbI can live with that.21:38
*** zehicle_at_dell has joined #openstack-infra21:39
anteayaSergeyLukjanov: okay so SlickNik is aware that SnowDust wants an account is is also aware if there are problems with the account, I will be talking with SlickNik and letting him deal with SnowDust21:39
*** arnaud has joined #openstack-infra21:40
*** david-lyle has quit IRC21:40
anteayaSergeyLukjanov: my only comment is that email uses and underscore in the username and I would like a hyphen used, but you can tweak that yourself21:40
*** david-lyle has joined #openstack-infra21:40
*** david-lyle has quit IRC21:40
*** pcm_ has left #openstack-infra21:41
*** pcm_ has joined #openstack-infra21:42
clarkbryanpetrello: when you calculate OG have you hardcoded a mash efficiency for all grain?21:42
pcm_dhellmann: ping21:42
*** sballe has joined #openstack-infra21:42
*** mestery has joined #openstack-infra21:42
*** pdmars has quit IRC21:42
*** jamielennox is now known as jamielennox|away21:43
*** zehicle_at_dell has quit IRC21:44
dhellmannpcm_: pong21:45
pcm_dhellmann: RE oslo.messaging -> requirements. Looks like it is still failing jenkins. Any ideas?
dhellmannpcm_: no, I'll have to look at that tomorrow21:45
mark0nI'm running into problems running Jenkins Job Builder on Debian Wheezy: "ImportError: No module named http_client". Installation with 'python install' completed without any problems. Any changes regarding the dependencies recently?21:45
pcm_dhellmann: OK thanks!21:46
SlickNikSergeyLukjanov: what anteaya said.21:48
SlickNikAlso thanks to both of you! :)21:48
jeblairmark0n: do you have a full traceback?21:48
ryanpetrelloclarkb: it’s configurable if you sign up21:49
*** rlandy has quit IRC21:50
ryanpetrellodefaults to .75, I think21:51
jeblairmark0n: if so, can you paste it into
*** mrda-away is now known as mrda21:51
jeblairmark0n: i think we just released a new python-jenkins that incorrectly omitted six from the runtime dependencies21:52
jeblairmark0n: if you "pip install six" (or equivalent) i think that should fix it21:53
jeblairmark0n: i expect a new python-jenkins release that corrects that soon21:53
*** praneshp has joined #openstack-infra21:53
nibalizermordred: why do you want puppetdb open to the world?21:54
nibalizerare you playing with puppet explorer?21:54
openstackgerritRamy Asselin proposed a change to openstack-infra/zuul: Remote push_change_refs from documentation
mark0njeblair: seems like you're right. But why does this affect my Debian Wheezy system (Wheezy is around for a while now)?21:55
*** Sukhdev has quit IRC21:55
openstackgerritRamy Asselin proposed a change to openstack-infra/zuul: Remove push_change_refs from documentation
zaroclarkb: missed the python-jenkins discussion.21:55
*** juzuluag has joined #openstack-infra21:55
mark0njeblair: my old six version was 1.1.021:55
jeblairmark0n: you appear to be installing jjb from source (which will install python-jenkins from source)21:56
jeblairor pip21:56
openstackgerritMichael Krotscheck proposed a change to openstack-infra/storyboard-webclient: Added StoryBoard administration module.
*** pballand has joined #openstack-infra21:56
mark0njeblair: I see21:56
openstackgerritMichael Krotscheck proposed a change to openstack-infra/storyboard-webclient: New StoryBoard Logo
openstackgerritMichael Krotscheck proposed a change to openstack-infra/storyboard-webclient: [WIP] Admin Project Group UI
zaroclarkb: should we backport your fix to python-jenkins before pbr and create a minor release on that?21:56
clarkbzaro: I don't think so. we can just release with pbr right?21:57
jeblairwhat would be the point of having made the release we just did?21:58
clarkbjeblair: oh this is the broken release thing?21:58
*** dizquierdo has joined #openstack-infra21:58
clarkbin that case yes we should backport ...21:58
*** maxbit has quit IRC21:59
zaroclarkb: yes, but wouldn't that mean the one without pbr is borken?22:00
*** maxbit has joined #openstack-infra22:00
mordrednibalizer: no, I was looking at using a rest call to puppetdb for the ansible inventory instead of puppet cert list22:00
mark0njeblair: thanks, now I have I workaround :-)22:00
zaroclarkb: ohh, you just said backport.22:00
mordrednibalizer: but that requires access from at least ci-puppetmaster ...22:00
mordrednibalizer: but .. I don't know whether that's a good or bad idea, or if there isa  reason to keep it locked down from the world22:01
zaroclarkb: so can we do a cherry-pick of your change onto 3.0-fixes branch?22:02
clarkbzaro: ya go for it, unless you want me to propose it22:02
zarowill do22:02
*** moted has quit IRC22:03
*** jay-house-hunter has joined #openstack-infra22:03
jeblairsdague: i think i have the quirks worked out now; this should work immediately on load as well as after navigating:
openstackgerritJames E. Blair proposed a change to openstack-infra/config: WIP: add CI results table
*** signed8bit is now known as signed8bit_ZZZzz22:04
*** zehicle_at_dell has joined #openstack-infra22:04
*** Sukhdev has joined #openstack-infra22:04
*** signed8bit_ZZZzz is now known as signed8bit22:04
sdaguejeblair: yep, seems to. Thanks!22:05
openstackgerritJames E. Blair proposed a change to openstack-infra/config: WIP: add CI results table
openstackgerritK Jonathan Harker proposed a change to openstack-infra/config: Make zuul, gerrit, and recheckwatch storyboard-aware
*** maxbit has quit IRC22:09
*** mriedem has quit IRC22:09
*** emagana has joined #openstack-infra22:09
*** maxbit has joined #openstack-infra22:09
asselinclarkb, how do you access the zuul refs via http?
emaganaHi! One test failed with the following error. I am not modifying any of the lib dependencies22:10
clarkbasselin: we setup apache in front of those git repos22:11
asselinclarkb, I have the portions in my apache that are in the docs.
mark0njeblair: can you point me to the bug report/fix for this issue?22:11
mark0njeblair: so that I can track when it made it into Debian?22:12
asselinclarkb, but i don't know how to connect the dots from http to what's on then filesystem/directives22:12
clarkbasselin: so you would do git fetch http://ip/openstack/foo master/Zf596c10ec43f40bc91c5e5519170b26022:12
*** miqui_ has joined #openstack-infra22:13
*** miqui has quit IRC22:13
asselinclarkb, in the browser, should this return anything?  http://ip/openstack/foo22:14
clarkbasselin: no22:14
clarkbasselin: it uses the git http protocol which isn't quite something a browser speaks22:15
anteayamorning mattoliverau22:15
*** maxbit has quit IRC22:16
anteayaemagana: so it looks like horizon requires django-pyscss>=1.0.1 and it either isn't on pypi.o.o or pypi.o.o couldn't be reached22:16
*** maxbit has joined #openstack-infra22:16
anteayadid we have pypi.o.o issues at all today?22:16
anteayaI haven't read all the backscroll22:16
*** datsun180b has quit IRC22:17
*** maxbit has quit IRC22:21
*** maxbit has joined #openstack-infra22:21
anteayabackscroll is indicating it *might* be an issue with pypi.o.o but I haven't come across a bug number yet22:22
clarkbanteaya: hpcloud has had trouble talking to it at times22:23
clarkbanteaya: the package is there22:23
clarkbemagana: anteaya the full log would be useful22:24
emaganaclarkb: let me get it22:24
clarkbhuh the error does talk about using allow external22:25
*** weshay has quit IRC22:25
clarkbthat is really weird as the tarballs exist on our mirror22:25
*** prad has quit IRC22:26
clarkbbut the test did run on hpcloud so maybe it is related to that trouble?22:26
*** duran has quit IRC22:27
clarkbdstufft: what does that real name of django_pyscss is django-pyscss mean?22:27
clarkbdstufft: is pip running into a weird requirements listing or something like that?22:27
dstufftclarkb: so packaging considers some characters equivilant22:29
*** eharney has quit IRC22:29
clarkbdstufft: oh ok so its doing a translation22:30
*** doug-fish has left #openstack-infra22:30
dstufftso the "name" of that project is django-pyscss, that's what it's registered with on PyPI22:30
clarkbdstufft: second question exists and has tarballs22:30
dstufftbut someone depended on django_pyscss somewhere22:30
clarkbwhy does it talk about external links22:30
*** Evgeny has joined #openstack-infra22:30
dstufftbecause you're using a dumb mirror22:30
dstufftmore or less22:30
dstufftsec gotta let dog out and then i'll explain22:30
clarkbdstufft: thanks22:31
*** sweston has quit IRC22:31
dstufftclarkb: ok22:31
dstufftclarkb: so basically, pip doesn't know what the "real" name of something is without determining it from PyPI, so if you depend on django_pyscss, it'll hit /simple/django_psyscss/VERSION/ (if you have a ==), if that 404's it'll hit /simple/django/psycss/, if that 404's it'll hit /simple/ and look for something that is named equivilantly22:33
*** maxbit has quit IRC22:33
*** weshay has joined #openstack-infra22:33
*** weshay has quit IRC22:33
*** reed has quit IRC22:33
dstufftthis works fine on PyPI, because PyPI is "smart" and it'll redirect /simple/django_psyscss/ to /simple/django-psyscss/ and it just works22:33
*** sweston has joined #openstack-infra22:34
dstufftbut you have a "dumb" mirror (just static files being served by apache), so /simple/djnago_psyscss/ 404's, causing pip to pull down /simple/ and look22:34
dstufftand apparently that confuses our external link detection code22:34
bodepdI started getting this error from jjb today:
clarkbdstufft: ok so it should still work with our dumb server, except that simple/ is huge and probably didn't load completely or something22:35
dstufftclarkb: I bet it just got confused because the /simple/ links to /simple/django-psyscss without an rel="internal" so pip sees the link to /simple/django-psyscss./ as external22:36
dstufftpip that is22:36
dstufftthis probably isn't well tested because most people don't bother with mirrors22:36
*** zehicle_at_dell has quit IRC22:36
* mordred cries22:36
dstufftclarkb: remedies -> use the real name, or use a tiny little web app instead of a pure dumb mirror22:37
mark0nclarkb: where do I find your change that got merged with ? I cannot find it on ...22:37
clarkbdstufft: ok22:37
mordreddstufft: we can't always use the real name22:37
clarkbemagana: so you want to change any django_pyscss to django-pyscss22:37
mordreddstufft: because of transitive depends22:37
dstufftso tiny web app then22:37
mordredbut I do agree - we shoudl absolutely use correct names in our requirements files22:37
emaganaclarkb: no, my change does not change anything there22:37
asselinclarkb, sorry still now working. any other ways to test/debug? git fetch http://localhost/python-3parclient master/Z72d2c156e1424a01b5aabad8774ec73a22:37
asselinfatal: http://localhost/python-3parclient/info/refs not found: did you run git update-server-info on the server?22:37
mordreddstufft: I'm going to lose my mind22:37
clarkbemagana: right its transitive probably22:37
*** boris-42 has quit IRC22:37
dstufftmordred: sorry22:37
*** jay-house-hunter has quit IRC22:38
clarkbemagana: or maybe horizon already depends on that using the wrong name and we just need to change it22:38
clarkbin any case the idea that _ and - are equivalent is silly22:38
*** zehicle_at_dell has joined #openstack-infra22:38
dstufftthey are equiv22:38
dstufftin packaging22:38
*** amalagon has quit IRC22:38
asselinclarkb, ls /var/lib/zuul/git/python-3parclient/22:38
asselindocs  hp3parclient  LICENSE.txt  README.rst  samples  test  zuulTest22:38
emaganaclarkb: that sounds like what is hapenning22:38
dstufftyou can register django-psyscss and django_psyscss at the same time22:38
mordredjeblair: ^^ check it out - we need to run a web app to serve files22:38
*** boris-42 has joined #openstack-infra22:39
clarkbasselin: check your apache logs?22:39
dstufftmordred: you only need to serve /simple/<foo>/, the files themselves can be pure static and so can /simple/22:39
dstufftyou just need /simple/<foo>/ to respond even if case differs22:39
clarkbwhat if we mod rewrite to catch 404s like that22:39
clarkbis that possible?22:39
*** whoops has quit IRC22:40
dstufftI forget what is possible with mod_rewrite22:40
dstufftbut we don't normalize the "stored" names, they are stored as whatever the author picked22:40
dstufftso you'll have some like foo-bar and some like foo_bar22:40
mark0nclarkb: but the gerrit reports the status as "merged"22:41
clarkbmark0n: we don't host the code for that project on launchpad anymore22:41
*** thedodd has quit IRC22:41
clarkbmark0n: its possible work needs to be done on launchpad to make that clear. You can git clone
dstufftckarre.sub('[^A-Za-z0-9.]+', '-', name)22:42
dstufftclarkb: re.sub('[^A-Za-z0-9.]+', '-', name)22:42
dstufftthat's what we run on names to normalize them22:43
dstufftwhen you go to /simple/<something whose exact name does not exist>/22:43
mark0nclarkb: I see. Thanks22:43
mordreddstufft: in pip?22:43
dstufftmordred: no PyPI22:43
dstufftthat's the behavior of /simple/<foo>/ in PyPI if /simple/<exact name>/ doesn't exist22:43
dstufftwe normalize the left and right hand size with that regex and look for a match22:44
asselinclarkb, I restarted apache and got these errors. I'm using jaypipe's third party ci setup.
clarkbemagana: you want to change that line22:44
*** praneshp has quit IRC22:44
*** moted has joined #openstack-infra22:44
dstufftmordred: clarkb sorry this kinda sucks right now, mirroring isn't the highest priority (but still important) and this just didn't get noticed22:45
dstufftand it's fallout from trying to kill externals22:45
clarkbemagana: you might want to leave a comment there that it really is a -22:45
clarkbsince the other django packages use _22:45
ryanpetrelloyou all aware of any issues w/ ceilometer in the gate?22:46
ryanpetrellothis is passing for me just fine in check, but when it goes to merge, it just keeps timing out22:46
mordreddstufft: sok. I'm just grumbly because I need to go write a web app now22:46
clarkbmordred: what if you enact the change in the world that we need or however that goes22:46
*** praneshp has joined #openstack-infra22:47
clarkbmordred: eg our mirror is "correct" please fix names if they are broken as in this case22:47
mordredclarkb: transitive deps22:47
mordredclarkb: over which we have zero control22:47
clarkbmordred: well we either remove the dependency that is broken or fix it22:47
mordredunless we start putting transitives earlier in our own files to get them installed that way22:47
clarkbwe can do that too I suppose22:48
emaganaclarkb: so, do I need to file a separate patch for this?22:48
clarkbemagana: I would yes22:48
mordreddstufft: would pip accept a patch to apply the normalization logic as the first attempt at a retry?22:48
emaganaclarkb: just wondering why other patches are not hitting this, are they?22:48
dstufftmordred: it won't work22:49
dstufftmordred: we don't know what form the name is on PyPI22:49
*** maxbit has joined #openstack-infra22:49
mordredoh. right22:49
clarkbemagana: its a good question. horizon is clearly working mostly22:49
clarkbemagana: so I am not sure why it broke here but not always22:49
emaganaclarkb: just don't want to break something that is working already  :-)22:49
*** dstanek is now known as dstanek_zzz22:49
zarojeblair, clarkb: for minor release of python-jenkins AND
clarkbemagana: changing that will only make it work more22:49
dstufftif PyPI normalized the /simple/<foo>/ instead of using whatever name the author used then it'd work22:49
dstufftwhich is what we should do in conjunction with that22:50
clarkbdstufft: what if pip failed, normalized and tried again22:50
clarkbdstufft: worst case it does a few more round trips and fails out22:50
emaganaclarkb: ok, I will do that22:50
dstufftclarkb: it'll only work if you get lucky22:50
clarkbmordred: jeblair any chance that can get through today? will make it possible to dig into logstash more later22:51
clarkbeither at the meetup or tomorrow/friday22:51
mordredclarkb: looking22:51
mordreddstufft: nod22:51
dstufftmordred: I'm on painkillers again for more dental work I had today, so I'm having a little trouble thinking through the rammifications, I *think* with a minor change to PyPI we can make it work though22:51
dstufftbut I can't think through rammifications to existing mirrors22:52
dstufftI think we'll have to insert a bunch of journal entries and essentially invalidate a bunch of stuff that's already been downloaded on the mirrors22:52
mordreddstufft: yah. that'll be fun for people22:52
dstufftbecause we can make it so that PyPI applies normalization in the simple API, so that PyPI will have something consistent for /simple/<whatever>/22:52
dstufftthen pip can know ahead of time what form the <whatever> will be in22:53
*** andreaf has quit IRC22:53
mordredyah. that all makes sense - the invalidation journal entries I think is probably the craziest part, yeah?22:53
dstufftbut I'm pretty sure it'll require tricking mirrors into thinking that for any package which isn't already in that normalized form, that it was deleted and then reuploaded so that they delete all the old files, and redownload them22:54
mordredalso - yay painkillers!22:54
dstufftand maybe people will be mad if we invalidate a signifcant chunk of their mirrors?22:54
mordredwell, _I_ won't :)22:54
clarkbwhat if pypi stops normalizing >_>22:54
dstufftclarkb: then we break lots of shit22:54
clarkband we force people to use the correct name22:54
clarkbdstufft: well ya but sometimes...22:54
dstufftlike part of normalization is capitalization22:54
dstufftpip install django would fail22:55
dstufftand pip install flask22:55
*** andreykurilin_ has quit IRC22:55
dstufftyou'd have to do pip install Django22:55
dstufftand pip install Flask22:55
clarkbthats fine too. maybe do it during pycon next year22:55
dstufftI don't think the normalization is the problem, it's just that we don't do enough of it22:55
clarkbthen everyone can yell at each other for using silly capitalization22:55
clarkband _s22:55
dstufftit's like xml, if it's not the solution you're not using enough of it22:55
dstufftand violence22:55
clarkbdstufft: honestly foo_bar != Foo_bar != foo-bar !=foo-Bar22:55
clarkbbecause these are things that live on case sensitive filesystems22:56
clarkband case sensitive files systems treat all of those diffrent22:56
dstufftbut these are also things that live on case insensitive filesystems22:56
clarkbleast common denominator here is case sensitive22:56
*** mark0n has quit IRC22:56
*** msabramo has quit IRC22:57
clarkbwe should just get the world to agree that filesystems are case sensitive22:57
dstufftbut really it doesn't cause any problems except it prevents the use of a fully dumb mirror, and the only reason it presents a problem there is because we don't normalize all the time22:57
dstufftI don't think it's very user friendly to yell at someone and fail because they picked the wrong case22:57
*** otherwiseguy has quit IRC22:57
*** mestery has quit IRC22:57
*** camunoz_gone is now known as camunoz22:59
*** atiwari has quit IRC22:59
*** shashankhegde has quit IRC23:00
*** amalagon has joined #openstack-infra23:00
*** adalbas has quit IRC23:00
clarkbdstufft: I have zero problems with my case sensitive fs23:01
dstufftclarkb: I would suggestion that the constraints of a local FS are probably different than the constraints of a unique identifier for package names23:02
dstufftit's similar to suggesting that the domain name system should be case sensitive ;)23:02
*** shashankhegde has joined #openstack-infra23:02
*** rkukura has quit IRC23:02
*** shivharis has quit IRC23:03
dstufftregardless we're going to apply some normalization to keep people from registering django and Django both, so if we already consider those two things equivalent, not giving people the one they want is just being punitive because they picked the wrong "form"23:03
clarkbdstufft: I think the big difference here is that dns records aren't installed onto my fs23:03
clarkbpypi packages are23:03
*** marcoemorais has quit IRC23:03
openstackgerritMonty Taylor proposed a change to openstack-infra/config: Add tools to build dib images with docker
*** marcoemorais has joined #openstack-infra23:04
dstufftclarkb: PyPI project names aren't installed onto your FS either really ;), at most the .egg-info/dist-info is stored there, but pip/setuptools handles that just fine anyways23:04
*** marcoemorais has quit IRC23:04
jeblairthere's an entire rfc on dns case sensitivity; i'm pretty sure that's a red herring and we should just drop the subject.  :)23:04
*** marcoemorais has joined #openstack-infra23:04
*** signed8bit has quit IRC23:04
lifelessdstufft: it would be nice to have pip not have to download the entire index to resolve case differences23:05
lifelessdstufft: yes it works, but sheese its overhead23:05
emaganaclarkb: I filed this one:
uvirtbotLaunchpad bug 1339918 in horizon "Requirements should django should be with hyphen" [Undecided,New]23:05
lifelessdstufft: (and the server doing the same thing is just as much work only minus the latency aspect)23:05
dstufftlifeless: so that is inherited from setuptools, we can fix it by having PyPI *always* serve the normalized form instead of serving whatever the author input, then pip can just blindly normalize too23:06
mordredjeblair, clarkb: ^^ the above patch is mainly so I can share current dev state23:06
*** markwash has joined #openstack-infra23:06
dstufftthe only problem with doing that, is it's going to invalidate a bunch of people's already cached stuff23:06
dstufftthat's not an insolvable problem or anything23:06
dstufftand we should probably do it23:06
marunsdague: hi!23:06
jeblairmordred: cool!  i gave it a quick -1 because it has your homedir in it, but don't let that fool you into thinking i'm not enthusiastic!  :)23:06
lifelessdstufft: +123:06
dstufftI just can't do it right tonight because getting the exact correct journal entries to minimize that requires more thinking then I'm capable of atm23:07
*** msabramo has joined #openstack-infra23:07
JayFmordred: !!!!!!23:07
sdaguemarun: hey23:07
mordredjeblair: wait - why is my homedir no good for you?23:07
JayFmordred: I <3 docker and will review the crap outta that23:07
mordredJayF: well, there is very little there :)23:07
clarkbmordred: you should give us all access to your homedir23:07
JayFmordred: thanks a bunch, you know this exists?
marunsdague: Do you have a moment to talk about post_test_hook and sudo?23:07
mordredclarkb: I will - as soon as we have AFS, I'll grant system:anyuser r access23:08
sdaguemarun: I'm almost done slaying generators, give me about 10 minutes23:08
*** jergerber has quit IRC23:08
marunsdague: roger that, please ping when ready23:08
emaganaclarkb: patch is here:
clarkbemagana: you should probably note why the change is happening in the commit message23:09
mordredJayF: yes - that's a bit more than what we need for this  - but I actually used your Dockerfile as the source of cargo culting for mine23:09
clarkbemagana: but other than that +123:09
emaganaclarkb: Let me add that to the bug description23:09
dstufftlifeless: mordred If I don't remember please poke me about fixing the normalization hole we just talked about in a day or two?23:09
clarkbemagana: well it should be in the commit too23:10
clarkbemagana: commit messages should always say why23:10
*** zzelle__ has quit IRC23:10
mordredJayF: in this case, I'm mainly doing a quick poc of running dib inside of a docker container so that I can avoid accidentally screwing the mysql db on th host23:10
*** bknudson has quit IRC23:10
clarkbemagana: the what is told by git diff, the why goes in the message.23:10
JayFmordred: We're trying to cull that down some. Get to a point where you build a tar'd filesystem, manipulate build params to give it a remote cloud-config.yml, then have it fetch the tar'd filesystem container on the fly23:10
mordredJayF: but I do need to remember to bother you for feedback on thigns23:10
JayFmordred: I'll add my name to that for reviews, I'd love to look at it and help guide if I can23:10
*** sweston has quit IRC23:11
sdagueclarkb: you are either going to love me or hate me for this next patch23:11
clarkbsdague: uh oh23:11
emaganaclarkb: sounds good, will update the commit message23:11
lifelessdstufft: fix the normalisation hole we spoke about in a day or two :)23:12
jeblairmordred: your change (or more likely, a parent of your change) has a merge conflict23:13
*** shashankhegde has quit IRC23:13
mordredjeblair: nod23:13
mordredjeblair: I actually should split it into two changes23:13
mordredjeblair: because one can land right now and is directly and immediately useful for hacking on nodepool things using current nodepool scripts23:14
mordredthen, the other half of the change is a potentially helpful addition to the dib-nodepool patches23:14
*** gyee has quit IRC23:14
emaganaclarkb: done!23:15
*** dangers is now known as dangers_away23:16
*** pcm_ has quit IRC23:16
clarkbnow that mordred has pushed that /me bugs about again :)23:16
*** HenryG has quit IRC23:16
*** melwitt has quit IRC23:16
jeblairclarkb: looking23:17
*** emagana has quit IRC23:17
*** morganfainberg is now known as morganfainberg_Z23:17
jeblairclarkb: i'm guessing we can refactor those constants out somewhere else later if we like this, yeah?23:18
clarkbjeblair: yeah23:19
jeblairclarkb: +223:19
clarkbjeblair: the node.* will probably stay there but the others hsould be cnetralized and snyced with the reast of the cluster23:19
openstackgerritSean Dague proposed a change to openstack-infra/os-loganalyze: Move all presentation layer to view module
sdagueclarkb: ok, so that's the presentation move around23:19
sdaguewhich gets rid of the regexes23:19
sdaguewell the dup regexes23:19
*** Sukhdev has quit IRC23:21
marunsdague: brb23:21
*** marun has quit IRC23:21
openstackgerritJames E. Blair proposed a change to openstack-infra/config: Make /var/run/puppet in apply test
openstackgerritMonty Taylor proposed a change to openstack-infra/config: Add tools to hack on infra with Docker
mordredjeblair: there you go ^^ that one is potentially useful currently23:21
*** markmcclain has quit IRC23:21
*** obondarev has quit IRC23:22
sdagueclarkb: also, thank diety for unit tests23:23
sdagueif I hadn't written those, I'd have been completely borked on this refactor23:24
*** krtaylor_away is now known as krtaylor23:24
*** shashankhegde has joined #openstack-infra23:26
openstackgerritMonty Taylor proposed a change to openstack-infra/config: Add support for disk-image-builder in nodepool
openstackgerritMonty Taylor proposed a change to openstack-infra/config: Add pre-requisites needed for diskimage-builder
mordredjeblair, clarkb: okie. I split the patch in twain and rebased it all23:27
mordredclarkb: I still havent' fixed your comments in 88479 ...23:28
*** dkliban_brb is now known as dkliban_afk23:28
*** nelsnelson has quit IRC23:28
clarkbsdague: why did get_generator in the tests appear to change its return type?23:29
clarkbsdague: is it returning filter now which must be cast to iter()?23:29
sdaguebecause it's iterable23:30
sdaguebut not an iterator23:30
sdagueso basically, I need to do that to be able to use next()23:30
sdagueotherwise I have to implement next23:31
sdaguewhich... isn't straight forward23:31
sdagueI will fully accept there might be another better solution, it is the one I came up with23:31
*** emagana has joined #openstack-infra23:33
clarkbno its fine, just making sure I understood23:33
clarkbsdague: also return hax :P23:34
clarkbbreak almost seems better but meh23:34
sdagueoh, in the text tests?23:34
sdagueyeh, that was before I figured out the iter() thing23:35
sdagueI should change those to that instead23:35
openstackgerritMonty Taylor proposed a change to openstack-infra/config: Add support for disk-image-builder in nodepool
openstackgerritMonty Taylor proposed a change to openstack-infra/config: Add pre-requisites needed for diskimage-builder
*** che-arne has quit IRC23:36
*** dizquierdo has quit IRC23:36
mtreinishclarkb, fungi, jeblair: any thoughts on pushing this through?:
*** juzuluag has quit IRC23:36
mtreinishI'd like to have it in stackforge before Germany23:36
clarkbsdague: ok commented23:38
*** Sukhdev has joined #openstack-infra23:38
sdagueclarkb: actually, if I was less dumb and put that in def get_generator, that would have been smart23:38
openstackgerritMonty Taylor proposed a change to openstack-infra/config: Add support for disk-image-builder in nodepool
mordredsorry for the spam23:39
openstackgerritMonty Taylor proposed a change to openstack-infra/config: Add pre-requisites needed for diskimage-builder
openstackgerritMonty Taylor proposed a change to openstack-infra/config: Add tools to hack on infra with Docker
clarkbsdague: that works too23:39
*** packet has joined #openstack-infra23:40
* clarkb reviews the new-project list23:40
clarkbdhellmann: mordred for oslo projects are we adding new ones as
*** sabeen1 has quit IRC23:41
clarkbI thought we were giving up on python package namespaces or am I mistaken?23:41
clarkbdims_: ^23:41
mordredclarkb: I thought we were not using .'s23:41
jesusauruswhere does the new contributer gerrit comment come from? where can i propose a change to it?23:42
clarkbmordred: me too, I guess I will ask dims23:42
clarkbjesusaurus: which comment?23:42
jeblairjesusaurus: it's in infra/config23:42
clarkbjesusaurus: the thing that happens when you push your first change up?23:42
jesusaurusclarkb: "Welcome, new contributor!"23:42
clarkbyea its in infra/config23:43
pleia2jesusaurus: AY23:43
jesusaurusjeblair: i just git-grepped and didnt see it23:43
jesusaurusahh jeepyb23:43
jeblairwhat a strange place23:43
jesusaurusthanks pleia223:43
pleia2I was just looking around jeepyb earlier for a patch, stumbled upon it :)23:43
pleia2else I would have forgotten too23:43
clarkbhuh I thought the message was config23:44
clarkbI guess that was the work mikal did and not that change23:44
openstackgerritSean Dague proposed a change to openstack-infra/os-loganalyze: Move all presentation layer to view module
clarkbmtreinish: you need to rebase apparently23:45
clarkbmtreinish: thats what you get for adding your project at the end of the file :P23:45
mtreinishclarkb: heh, ok one sec23:45
jeblairsdague: is the approximate visual look i'm going for23:45
sdaguejeblair: ++23:45
jeblair(i think eventually it probably wants a patchset selector, but that can come later)23:46
sdagueyeh, probably23:46
sdaguehonestly, it also occurs to me that with lots of results, the fails are probably all that need to be shown at first glance, as long as there is an easy way to also find the passes.23:47
clarkbI don't know if the monasca folks are around but I think it was fuel that decided is an anti pattern23:47
clarkbmordred: ^23:47
clarkbI am fine with it I guess, but the attic is going to get full over time :)23:47
StevenKclarkb: So then we create openstack-storageshed and move stuff there?23:48
clarkbStevenK: eventually we will have a warehouse23:48
jeblairand a leopard23:48
StevenKjeblair: -and-the-stairs-are-missing23:48
*** jamielennox|away is now known as jamielennox23:49
jesusaurusStevenK: isnt that just -the-pit?23:49
openstackgerritMatthew Treinish proposed a change to openstack-infra/config: Add subunit2sql to stackforge
mtreinishclarkb: ^^^23:49
mtreinishall those xstatic projects...23:50
sdagueyeh, those xstatic projects are kind of an anti pattern as well23:50
*** marun has joined #openstack-infra23:51
marunsdague: back23:51
clarkbsdague: in more ways than one23:52
jeblairsdague: how so?  (not that i'm a fan...)23:52
jeblairclarkb: you too :)23:52
sdaguejeblair: it seems very weird to me to build a large number of pips on each for every js library out there23:53
*** emagana has quit IRC23:53
*** emagana has joined #openstack-infra23:53
clarkboh sorry it was murano23:54
marunsdague: post_test_hook is invoked after this:
*** shayneburgess has quit IRC23:54
sdagueand then maintain that, given that horizon is probably not going to test against every version of that matrix23:54
openstackgerritK Jonathan Harker proposed a change to openstack-infra/jeepyb: Add commit message link to newbie message
clarkbjeblair: that. I don't think js libs should be put into python packages23:54
sdaguemarun: ah, so root wrap doesn't help you23:55
marunsdague: nope.  it doesn't look like there is a way to run a sudo-requiring command with the devstack-vm-gate script23:56
marunsdague: any ideas?23:56
clarkbmarun: wait what? devsatck require sudo23:56
marunsdague: I can't see a way of running the neutron functional job with sudo without changing the devstack-gate scripts23:56
sdaguemarun: in your pre_hook drop your own sudoers.d file in there?23:56
clarkband it is run by devstack-gate23:56
clarkblets back up. what are we trying to do in the post_test_hook?23:57
marunclarkb: post_test_hook is run after devstack is run, after sudo is removed23:57
clarkbare we running tests there? can we move the tests to the test_hook?23:57
marunclarkb: run a sudo-requiring command23:57
*** UtahDave has joined #openstack-infra23:57
marunclarkb: maybe?23:57
mordredsdague: I agree with you23:58
marunclarkb: I was only using post_test_hook because it was suggested to me23:58
*** emagana has quit IRC23:58
*** froek has joined #openstack-infra23:58
sdaguemarun: is this a command that you can virtually invoke by calling neutron? Or was that kind of the point.23:58
mordredsdague: it seems to me that, if one is going to use a javascript library, using the javascript library tooling and framework would be the way to go23:58
*** msabramo has quit IRC23:58
*** penguinRaider has quit IRC23:58
marunsdague: wouldn't matter either way, the rootwrap invocation would be the same23:58
mordredsdague: but $distros_confused so we do it this way23:58
mordredsdague: I have NO IDEA why this way is better for the distros, mind you23:59
clarkbmordred: well more node is confused23:59
clarkbmordred: beacuse node is terrible software23:59
clarkbthey don't accomodate the distros at all23:59
mordredclarkb: and we're going to replace that with pypi?23:59
clarkbwhich isn't surprising considering the source23:59
clarkbmordred: its terrible in ways that directly give the distros hell23:59
mordredclarkb: I disagree that it's terrible software. I think that the people who write it are assholes and very very bad humans to talk to23:59
clarkbmordred: for example when heartbleed happened anyone terminating ssl in node had to recompile23:59

Generated by 2.14.0 by Marius Gedminas - find it at!