Monday, 2013-11-25

*** markwash has joined #openstack-infra00:01
clarkbanteaya: yes00:01
clarkbok on a real computer now, looking at etherpad stuff00:02
*** markwash has quit IRC00:04
fungigrr. my first tempest run ran into bug 1217734 so this may take a while00:06
uvirtbotLaunchpad bug 1217734 in python-keystoneclient "FAIL: setUpClass (tempest.api.compute.servers.test_server_rescue.ServerRescueTestXML Unauthorized) " [High,New]
*** senk has quit IRC00:08
*** senk1 has joined #openstack-infra00:08
clarkbthere are 900 revisions of that pad00:13
clarkbtrying to bisect that list now to find where the list of actions lives00:14
clarkbI could also grab all of them and deal with it that way00:14
*** herndon has joined #openstack-infra00:16
anteayathat is a lot of revisions00:17
anteayaI'm wondering why that history isn't available via the gui00:17
*** sarob has quit IRC00:18
*** jhesketh_ has quit IRC00:18
clarkbanteaya: because timeslider is derpy, There is a long standing bug open on it not working always on the github issue tracker00:19
*** sarob has joined #openstack-infra00:19
clarkblifeless: so, revision 894 is the revisions that reset the data00:19
*** fifieldt has joined #openstack-infra00:19
clarkblifeless: revision 893 does not have an actions list00:19
clarkblifeless: I can paste 893 on paste.o.o00:19
lifelessclarkb: can you reset it back to 893 or something?00:19
*** senk1 has quit IRC00:19
lifelessclarkb: a paste would be a good start00:20
*** sarob has quit IRC00:23
anteayaclarkb: ah00:24
clarkblifeless: revision 893. The only editing I did was the replace \n with actual newlines00:25
clarkboh derp, my vim display trailing spaces as dashes shows up in there too :/00:26
lifelessclarkb: LOL00:26
clarkbbut other than those two things, that text is original :)00:26
*** dcramer_ has quit IRC00:27
clarkblifeless: hope that helps00:28
lifelessI'l paste that back on the pad00:28
lifelessif you can reset it that would be better00:28
clarkbfor those following along, I used the getRevisionsCount API call to get an upper bound, then worked backward with getText until I found the replacement00:28
clarkblifeless: I think that adding a new revision with the old content is preferable. Mostly because the DB schema for etherpad is weird and I don't want to do writes on it right now :)00:29
clarkbbut also for the same reasons you make git revert commits00:29
*** milki has quit IRC00:30
lifelessclarkb: mmm, no, this isn't a distributed system.00:31
lifelessclarkb: all of the git reasons have to do with that.00:31
clarkblifeless: right but the history of the derp is valid as history00:31
lifelessnot really00:32
clarkbwhy not? it happened00:32
clarkb(and people noticed it happened)00:32
clarkbI think this sort of discussion is similar to "should rebase ever be used"00:32
clarkbdarcs says no, many git users say yes, no one agrees00:33
lifelessactually I think you'll find darcs' internals are much more rebase like than you think00:33
lifelessparticular around ejecting patches accidentally merged00:34
lifelessrebase on public branches has issues which are due to the behaviour of distributed systems00:34
lifelesswhich this isn't...00:34
clarkbI thought darcs refuses to provide a rebase operation through its UI00:35
clarkbas a user you cannot rewrite history with darcs00:35
lifelessas a darcs user history is somewhat subjective00:35
clarkboh looks like they are adding proper rebase00:36
*** milki has joined #openstack-infra00:36
clarkbadded earlier this year00:37
lifelessif you were to argue 'hey, don't delete misguided edits, undo them' - thats fine. But 'someone messed up the entire state by mistake' - its a bit of a bigger fixup :>00:37
lifelessmove on00:37
*** markwash has joined #openstack-infra00:41
*** cyril__ has quit IRC00:42
lifelessclarkb: that gets at some of the plumbing I'm offhandedly referring to00:42
lifeless"As a Darcs user you are freed from a lot of the artefacts of worrying about commit order.  Collaborating with people is just question of shuffling patches around, with no merge states, no rebases, way fewer spurious dependencies to worry about."00:43
lifelessclarkb: I think you'll find it a fun read00:44
clarkbya reading it now, so far quite interesting00:45
anteayaI'm enjoying it00:46
clarkbdarcs rebase shows up too00:47
*** amotoki has joined #openstack-infra00:47
*** Alssi has quit IRC00:52
*** matsuhashi has joined #openstack-infra01:03
*** senk has joined #openstack-infra01:04
*** herndon has quit IRC01:04
*** danger_f` has joined #openstack-infra01:06
*** Shrews_ has joined #openstack-infra01:06
*** danger_fo_away has quit IRC01:07
*** Shrews has quit IRC01:07
*** danger_f` has quit IRC01:07
*** danger_f` has joined #openstack-infra01:07
*** senk has quit IRC01:08
fungiugh. gotten this tempest failure twice in a row...
*** matsuhashi has quit IRC01:11
*** matsuhashi has joined #openstack-infra01:11
*** matsuhas_ has joined #openstack-infra01:15
*** matsuhashi has quit IRC01:16
*** nosnos has joined #openstack-infra01:18
*** rongze has joined #openstack-infra01:19
*** rongze has quit IRC01:24
*** senk has joined #openstack-infra01:26
*** sarob has joined #openstack-infra01:29
*** sandywalsh has quit IRC01:30
clarkbpossibly related to libvirt?01:32
*** sarob has quit IRC01:34
fungii don't know. that was on rackspace so i'm retrying on hpcloud now01:35
*** sjing has joined #openstack-infra01:35
fungiif i can repeat it there consistently, i'm going to try toggling versions of libvirt for comparison01:35
*** gfa has quit IRC01:36
clarkbanything related to server build fails is going to make me wary as that is what libvirt is for01:36
clarkbgood idea01:36
*** dstanek has quit IRC01:36
portanteclarkb: hello, do all the log files make it into logstash?01:37
clarkbnot all but most01:37
clarkbthe lisg01:37
clarkber the list is at openstack-infra/config/modules/openstack_project/files/logstash/somethingclient.yaml01:38
portanteI noticed that searching for the string: txdb6be9b4235d41cfabb9d-00529077a7, did not hit
portanteon that gate job, for some reason the proxy-server entries did not land it syslog.txt, only in the screen session01:38
*** senk has quit IRC01:39
clarkbswift proxy logs are not indexed as they dont log enough info and have no apparent structure01:39
portanteoh, can we get that changed?01:39
lifelessfungi: I have now answered in that PCI passthrough thread.01:39
clarkbthere is apache combined stuff + whatever01:39
portanteclarkb: we need that stuff logged01:39
portanteindexed, sorry01:39
clarkbportante: if swift switches to consistent log format with timestamps yes we can add it01:39
portanteI htink we are noticing two different things01:41
*** dstanek has joined #openstack-infra01:41
portante1st, there are not proxy-server log entries in syslog.txt01:41
clarkbnotmyname said I could do this by writing a middleware iirc, my argument that seift should log cleanly by default lost01:41
portanteclarkb: I think that is a problem, as swift should be configured the same way as the object, container and account servers, which are logging to syslog.txt01:42
*** bingbu has joined #openstack-infra01:42
clarkb(bbackward compat being the consideration)01:42
portantecan I find the final proxy-server.conf file that is generated for a given run?01:42
clarkbyes it is in the logs dir on logs.o.o for that run01:43
clarkber no01:43
clarkbsorry thought you wanted the log file01:43
portanteI want to find out why the proxy-server log entries are not there for the grenade runs, I have seen them there in other runs01:43
clarkbthis is another issue with swift :) you can't move the conf dir which makes taring all the conf files to save them a pain01:44
clarkbother projects have this problem too01:44
portantethat would be a useful thing to support, I agree01:44
portanteI could then run comparison tests for performance much easier01:44
portanteso the /etc/swift/* directory is not saved after a run?01:45
*** guohliu has joined #openstack-infra01:45
clarkbit is not01:46
clarkbwe probably wont do that until we can have an /etc/devstack/* then tar all of /etc/devstack01:46
*** odyi has quit IRC01:46
clarkbwhat you can do is apply the transformations locally01:47
clarkbdevstack takes the in project example configs and updates them using functions in devstack/lib/$project01:47
clarkbif you grab that lib file you should be able to apply just the transformations without a full devstack run01:48
portanteclarkb, is devstack used for all check and gate jobs?01:51
*** sjing has quit IRC01:52
*** sjing has joined #openstack-infra01:53
*** yaguang has joined #openstack-infra01:53
clarkbfor any with devstack in the job name. grenade doesn't use devstack directly but does use the devstack/lib/project files01:53
clarkboh and swift functional uses devstack to install and start swift01:54
*** rongze has joined #openstack-infra01:56
*** rongze has quit IRC01:57
*** rongze has joined #openstack-infra01:57
*** rongze_ has joined #openstack-infra01:59
*** dcramer_ has joined #openstack-infra02:01
*** rongze has quit IRC02:02
*** paul-- has quit IRC02:06
openstackgerritNoorul Islam K M proposed a change to openstack-dev/cookiecutter: Add missing HACKING.rst
*** sarob has joined #openstack-infra02:16
*** CaptTofu has quit IRC02:20
*** CaptTofu has joined #openstack-infra02:21
*** paul-- has joined #openstack-infra02:24
*** dstanek has quit IRC02:35
openstackgerritNoorul Islam K M proposed a change to openstack-dev/cookiecutter: Add reasons to ignore certain flake8 checks
*** sarob has quit IRC02:52
*** gfa has joined #openstack-infra02:52
*** sarob has joined #openstack-infra02:53
*** gfa has left #openstack-infra02:53
*** gfa has joined #openstack-infra02:53
*** sarob has quit IRC02:57
*** rcleere has joined #openstack-infra02:57
*** sjing has quit IRC03:01
*** sjing has joined #openstack-infra03:03
*** paul-- has quit IRC03:08
*** rongze_ has quit IRC03:09
lifelessjog0: is E126 new?03:16
lifelessjog0: where do these new nits get discussed, cause I think E126 generates terrible formatting03:16
clarkbE126 is not new and from pep8, it is E125's siblif (elif insteaf of if iirc)03:16
*** matsuhas_ has quit IRC03:17
clarkband most agree with you about both03:17
lifelessit wants03:17
lifeless        help='Where to write output. If - stdout is used and only one output'03:17
lifeless            'format may be given. Otherwise the output format is appended to'03:17
lifeless        help='Where to write output. If - stdout is used and only one output'03:17
lifeless        'format may be given. Otherwise the output format is appended to'03:17
dstufftor it wants you to line up the '03:17
clarkbwe ignore both in many infra projects03:17
dstufftI think03:18
Alex_GaynorI don't think that's necessarily what it wants, there are definitely other ways to make it happy, I use this one and I find the latter supper ugly :-)03:18
dstufftand yea i find the latter ugly as sin too and I'm pretty sure I don't disable E126 anywhere03:18
lifelessdstufft: thats a 'visual indent' where it gives E128 if it's wrong.03:18
*** matsuhashi has joined #openstack-infra03:19
*** rongze has joined #openstack-infra03:23
openstackgerritlifeless proposed a change to openstack-infra/reviewstats: Tweak formatting for hacking.
openstackgerritlifeless proposed a change to openstack-infra/reviewstats: Cookiecutterify reviewstats.
openstackgerritlifeless proposed a change to openstack-infra/reviewstats: Ignore E126.
*** SergeyLukjanov has joined #openstack-infra03:39
*** dstanek has joined #openstack-infra03:40
*** matsuhashi has quit IRC03:40
*** dstanek has quit IRC03:41
*** mriedem has quit IRC03:46
*** chandankumar has quit IRC03:46
*** chandankumar has joined #openstack-infra03:48
*** CaptTofu has quit IRC04:03
*** sarob has joined #openstack-infra04:03
*** CaptTofu has joined #openstack-infra04:04
*** dkliban has joined #openstack-infra04:06
*** sarob has quit IRC04:07
*** oubiwann has joined #openstack-infra04:14
*** boris-42 has joined #openstack-infra04:24
*** masayukig has joined #openstack-infra04:26
*** yaguang has quit IRC04:27
*** markwash has quit IRC04:27
*** nicedice has quit IRC04:27
*** yaguang has joined #openstack-infra04:30
fungiclarkb: yeah, not going well. same server in error state failures on hpcloud as well (tried twice there). without enabling cloud archive, full tempest completes with no problems. i've followed up to the ml thread04:32
clarkbthank you04:34
jog0clarkb: ping04:36
jog0clarkb: can you review the wrapup at the bottom
jog0want to send it out today before the full monday deluge04:37
clarkbjog0 lgtm04:41
clarkbI made on spelling fix04:41
*** markwash has joined #openstack-infra04:41
jog0clarkb: cool04:42
jog0you think the email is on target etc?04:42
clarkbI think so. It explains what and why and our future plans04:42
jog0cool, off it goes04:43
jog0clarkb: can I put your name at the bottom04:44
jog0as you wrote a bunch of it04:44
clarkbonly if you put yours too :)04:44
jog0haha of course04:44
mordredholy shit sunday backscroll04:45
openstackgerritNoorul Islam K M proposed a change to openstack-infra/config: Make python33 gate voting for solum project
*** herndon has joined #openstack-infra04:47
clarkbmordred: just for you04:50
*** rongze has quit IRC04:50
jog0clarkb: hows this for the TL;DR TL;DR Last week the gate got wedged on nondeterministic failures, we took drastic actions and unwedged the gate.04:51
*** matsuhashi has joined #openstack-infra04:51
jog0mordred: ^04:52
*** SergeyLukjanov has quit IRC04:52
clarkb... took drastic actions to fix bugs unwedging the gate04:52
openstackgerritMichael Still proposed a change to openstack-infra/elastic-recheck: Tell people to do a recheck
*** rongze has joined #openstack-infra04:53
openstackgerritYuuichi Fujioka proposed a change to openstack-dev/hacking: Add metaclass for Python3 compatibility
*** amotoki has quit IRC04:55
*** amotoki has joined #openstack-infra04:55
mordredAlex_Gaynor, dstufft: the E12* checks in pep8 are super derpy and (saying this as a person who regularly writes code which overreaches authority) overreach the purpose of the tool04:56
jog0clarkb: TL;DR Last week the gate got wedged on nondeterministic failures. Unwedging the gate required drastic actions to fix bugs.04:56
jog0hows that04:56
mordredE123, E125 and E126 are not enforcements of PEP8, they are enforcements of random thoughts that the pep8 tool author had - which would be fine if they were not labeled pep804:57
* mordred likes E125 for instance04:57
mikaljog0: speaking of which, lifeless and I had a quick look at that create_backup test late last week04:57
mikaljog0: and then I decieded to go to dinner instead04:57
mordredmikal: dinner is always the right choice04:57
jog0mordred: sounds like a wise choice to me04:58
mikaljog0: but it looks like taking a snapshot in libvirt 0.9.8 is a cold process involving a new domain. That test does that over and over. Perhaps it somehow breaks the libvirt domain we're trying to get console logs from?04:58
mordredmikal: I will say that a possible argument against dinner is "booze" - but other than that, dinner should win04:58
mikalmordred: LOL04:58
clarkbjog0 lgtm04:59
jog0mikal: hmm when did we go to libvirt 0.9.804:59
clarkbprecise has 0.9.804:59
clarkband cloud archive libvirt appears broken04:59
jog0because when that teest passed the check queue every time but failed in gate and needed a reverify05:00
jog0anyway that bug isn't closed so  fill out the bug report with your findings if you haven't done so lalready05:00
jog0clarkb: email sent, mikal its still the weekend here so ttyl05:01
*** sjing has quit IRC05:05
*** oubiwann has quit IRC05:09
*** oubiwann has joined #openstack-infra05:09
zaroclarkb: you gonna be in tomorrow?05:15
clarkbzaro: no going to work remotely05:17
*** sarob has joined #openstack-infra05:22
*** sjing has joined #openstack-infra05:22
*** guohliu has quit IRC05:24
*** guohliu has joined #openstack-infra05:25
*** rpodolyaka has joined #openstack-infra05:27
*** jorisroovers has joined #openstack-infra05:35
*** sarob has quit IRC05:50
*** nosnos_ has joined #openstack-infra05:51
*** sarob has joined #openstack-infra05:51
*** herndon has quit IRC05:51
*** nosnos has quit IRC05:54
*** matsuhashi has quit IRC05:55
*** sarob has quit IRC05:55
*** nosnos has joined #openstack-infra05:57
*** nosnos_ has quit IRC05:57
*** matsuhashi has joined #openstack-infra05:59
*** sarob has joined #openstack-infra06:03
*** matsuhashi has quit IRC06:04
*** sjing has quit IRC06:07
*** matsuhashi has joined #openstack-infra06:19
*** rcleere has quit IRC06:32
*** jcoufal has joined #openstack-infra06:39
*** matsuhashi has quit IRC06:47
*** boris-42 has quit IRC06:47
*** rpodolyaka has left #openstack-infra06:50
openstackgerritlifeless proposed a change to openstack-infra/reviewstats: Move tests into the setuptools structure.
*** markwash has quit IRC06:57
*** matsuhashi has joined #openstack-infra06:58
*** matsuhashi has quit IRC06:59
*** matsuhashi has joined #openstack-infra07:02
*** D30 has joined #openstack-infra07:04
*** nosnos has quit IRC07:05
*** nosnos_ has joined #openstack-infra07:05
*** markwash has joined #openstack-infra07:11
openstackgerritLiang Chen proposed a change to openstack-dev/pbr: Make sphinx builders configurable in LocalBuildDoc
openstackgerritlifeless proposed a change to openstack-infra/reviewstats: Move utils into the reviewstats namespace.
openstackgerritlifeless proposed a change to openstack-infra/reviewstats: Basic usage instructions.
openstackgerritlifeless proposed a change to openstack-infra/reviewstats: Move openapproved into the reviewstats namespace.
openstackgerritlifeless proposed a change to openstack-infra/reviewstats: Move reviews_for_bugs into reviewstats.
openstackgerritlifeless proposed a change to openstack-infra/reviewstats: Move openreviews into the reviewstats namespace.
openstackgerritlifeless proposed a change to openstack-infra/reviewstats: Move into the reviewstats namespace.
*** matsuhashi has quit IRC07:23
*** matsuhashi has joined #openstack-infra07:23
openstackgerritlifeless proposed a change to openstack-infra/config: Document what it takes to be a check/gate test.
*** matsuhashi has quit IRC07:28
*** yolanda has joined #openstack-infra07:29
*** matsuhashi has joined #openstack-infra07:29
*** matsuhashi has quit IRC07:29
*** matsuhashi has joined #openstack-infra07:29
openstackgerritlifeless proposed a change to openstack-infra/config: Start running regular jobs on reviewstats.
*** dizquierdo has joined #openstack-infra07:32
openstackgerritlifeless proposed a change to openstack-infra/config: Start running regular jobs on reviewstats.
*** sjing has joined #openstack-infra07:34
*** sarob has quit IRC07:39
*** sarob has joined #openstack-infra07:39
*** afazekas has joined #openstack-infra07:41
*** _afazekas has quit IRC07:43
*** sarob has quit IRC07:43
lifelessjeblair: I am sure you have a tonne of backlog, but I'd love to have get an eye cast over it: I'd like to be able to point folk at reference docs for this ;)07:44
*** pblaho has joined #openstack-infra07:50
*** michchap has quit IRC07:53
*** michchap has joined #openstack-infra07:53
*** oubiwann has quit IRC07:57
*** flaper87|afk is now known as flaper8708:02
*** osanchez has joined #openstack-infra08:03
*** matsuhashi has quit IRC08:09
*** mattymo has quit IRC08:09
*** matsuhashi has joined #openstack-infra08:10
*** hashar has joined #openstack-infra08:10
*** dizquierdo has quit IRC08:10
*** plomakin has quit IRC08:11
*** dizquierdo has joined #openstack-infra08:14
*** nosnos_ has quit IRC08:14
*** sjing has quit IRC08:15
*** nosnos has joined #openstack-infra08:15
*** sjing has joined #openstack-infra08:16
*** hashar is now known as hasharCall08:23
*** SergeyLukjanov has joined #openstack-infra08:36
*** sarob has joined #openstack-infra08:40
*** sjing has quit IRC08:43
*** DinaBelova has joined #openstack-infra08:52
*** davidhadas has joined #openstack-infra08:56
*** paul-- has joined #openstack-infra08:56
*** derekh has joined #openstack-infra08:58
*** Grishkin has joined #openstack-infra09:01
*** masayukig has quit IRC09:05
*** jpich has joined #openstack-infra09:06
*** dizquierdo has quit IRC09:12
*** matsuhashi has quit IRC09:12
*** sarob has quit IRC09:13
*** matsuhashi has joined #openstack-infra09:14
*** yassine has joined #openstack-infra09:16
*** fbo_away is now known as fbo09:24
*** D30 has quit IRC09:25
*** markmc has joined #openstack-infra09:25
*** bingbu has quit IRC09:32
*** mihgen has joined #openstack-infra09:37
*** boris-42 has joined #openstack-infra09:42
*** jake2 has joined #openstack-infra09:45
*** psedlak has quit IRC09:45
*** SergeyLukjanov is now known as _SergeyLukjanov09:46
*** _SergeyLukjanov has quit IRC09:46
*** cyril__ has joined #openstack-infra09:48
*** SergeyLukjanov has joined #openstack-infra09:48
*** paul-- has quit IRC09:49
*** sarob has joined #openstack-infra09:49
*** jake2 has left #openstack-infra09:50
*** sarob has quit IRC09:54
*** marun has joined #openstack-infra10:00
*** paul-- has joined #openstack-infra10:02
*** salv-orlando has quit IRC10:05
*** salv-orlando has joined #openstack-infra10:05
*** rongze has quit IRC10:13
*** psedlak has joined #openstack-infra10:19
openstackgerritJaroslav Henner proposed a change to openstack-infra/jenkins-job-builder: Add batch_tasks support.
openstackgerritJaroslav Henner proposed a change to openstack-infra/jenkins-job-builder: Add seealso to batch_tasks from promoted_build.
*** pcm_ has joined #openstack-infra10:33
*** pcm_ has quit IRC10:34
*** pcm_ has joined #openstack-infra10:35
*** hdd has joined #openstack-infra10:40
*** sarob has joined #openstack-infra10:40
*** lcestari has joined #openstack-infra10:43
*** hdd has quit IRC10:45
*** paul-- has quit IRC10:46
*** thomasbiege1 has joined #openstack-infra10:49
*** dizquierdo has joined #openstack-infra10:53
openstackgerritA change was merged to openstack/requirements: Updated from global requirements
openstackgerritA change was merged to openstack/requirements: Add both climate repos for auto reqs update
*** yaguang has quit IRC10:54
*** paul-- has joined #openstack-infra10:56
*** ArxCruz has joined #openstack-infra10:58
*** matsuhashi has quit IRC10:58
*** boris-42_ has joined #openstack-infra10:59
*** plomakin has joined #openstack-infra11:00
*** boris-42 has quit IRC11:01
*** ruhe has joined #openstack-infra11:05
*** nosnos has quit IRC11:07
*** SergeyLukjanov is now known as _SergeyLukjanov11:11
*** thomasbiege1 has quit IRC11:12
*** wenlock has quit IRC11:13
*** rongze has joined #openstack-infra11:13
*** sarob has quit IRC11:14
*** adalbas has joined #openstack-infra11:14
*** rongze has quit IRC11:18
*** yongli has quit IRC11:20
*** yongli has joined #openstack-infra11:20
*** guohliu has quit IRC11:24
*** rfolco has joined #openstack-infra11:28
*** markmcclain has quit IRC11:30
*** zoresvit has joined #openstack-infra11:30
*** paul-- has quit IRC11:35
*** paul-- has joined #openstack-infra11:36
*** sarob has joined #openstack-infra11:40
*** SergeyLukjanov has joined #openstack-infra11:42
*** sarob has quit IRC11:45
*** ArxCruz has quit IRC11:51
*** DennyZhang has joined #openstack-infra11:54
*** ArxCruz has joined #openstack-infra11:55
*** boris-42 has joined #openstack-infra11:56
*** boris-42_ has quit IRC11:57
*** SergeyLukjanov is now known as _SergeyLukjanov11:59
*** _SergeyLukjanov is now known as SergeyLukjanov11:59
*** boris-42_ has joined #openstack-infra12:01
*** boris-42 has quit IRC12:03
*** che-arne has joined #openstack-infra12:07
*** rongze has joined #openstack-infra12:14
*** hasharCall has quit IRC12:16
*** michchap has quit IRC12:18
*** rongze has quit IRC12:18
*** michchap has joined #openstack-infra12:20
openstackgerritArx Cruz proposed a change to openstack-infra/zuul: Add the possibility to pass the zuul url
*** fifieldt has quit IRC12:22
*** amotoki has quit IRC12:23
*** rongze has joined #openstack-infra12:24
*** DinaBelova has quit IRC12:25
*** DinaBelova has joined #openstack-infra12:30
*** sarob has joined #openstack-infra12:40
*** davidhadas_ has joined #openstack-infra12:45
*** SergeyLukjanov is now known as _SergeyLukjanov12:46
*** ilyashakhat has joined #openstack-infra12:47
*** _SergeyLukjanov has quit IRC12:47
*** davidhadas has quit IRC12:47
*** yaguang has joined #openstack-infra12:51
*** hashar has joined #openstack-infra12:52
*** xchu has joined #openstack-infra12:52
chandankumarI am not able to login into openstack wiki using launchapd id12:56
chandankumarplease someone help me?12:56
*** SergeyLukjanov has joined #openstack-infra12:57
*** hashar has quit IRC12:59
*** dkranz has quit IRC13:00
*** dstanek has joined #openstack-infra13:00
*** paul-- has quit IRC13:02
*** xchu has quit IRC13:02
*** xchu has joined #openstack-infra13:04
*** paul-- has joined #openstack-infra13:05
*** odyssey4me has joined #openstack-infra13:11
*** hashar has joined #openstack-infra13:12
*** sarob has quit IRC13:13
*** davidhadas has joined #openstack-infra13:13
*** davidhadas_ has quit IRC13:17
*** sandywalsh has joined #openstack-infra13:17
*** bpokorny has quit IRC13:19
*** weshay has joined #openstack-infra13:20
*** ruhe has quit IRC13:21
*** yaguang has quit IRC13:22
*** thomasem has joined #openstack-infra13:22
*** thomasem has quit IRC13:23
*** thomasem has joined #openstack-infra13:23
*** rongze has quit IRC13:25
*** beagles has joined #openstack-infra13:27
*** nsaje has joined #openstack-infra13:28
*** dcramer_ has quit IRC13:31
*** yaguang has joined #openstack-infra13:35
*** ruhe has joined #openstack-infra13:37
*** sarob has joined #openstack-infra13:40
*** bpokorny has joined #openstack-infra13:41
*** boris-42_ is now known as boris-4213:41
*** paul-- has quit IRC13:43
*** sarob has quit IRC13:44
*** alexpilotti has joined #openstack-infra13:45
*** hdd has joined #openstack-infra13:46
*** herndon has joined #openstack-infra13:48
sdagueman, quieter monday morning than I'm used to :)13:48
krtaylormorning sdague13:49
mordredsdague: things don't seem to have fallen apart in the last couple of days13:52
sdaguenice, that's good to hear :)13:52
*** dkranz has joined #openstack-infra13:54
*** CaptTofu has quit IRC14:00
*** CaptTofu has joined #openstack-infra14:00
*** dprince has joined #openstack-infra14:01
*** DinaBelova has quit IRC14:02
*** yaguang has quit IRC14:02
*** DennyZhang has quit IRC14:03
*** rongze has joined #openstack-infra14:04
*** ewindisch has joined #openstack-infra14:05
*** hdd has quit IRC14:05
*** odyssey4me has quit IRC14:09
openstackgerritA change was merged to openstack-dev/cookiecutter: Add missing HACKING.rst
*** julim has joined #openstack-infra14:13
*** julim has quit IRC14:14
*** mriedem has joined #openstack-infra14:14
*** lascii is now known as alaski14:15
openstackgerritA change was merged to openstack-dev/cookiecutter: Add reasons to ignore certain flake8 checks
*** julim has joined #openstack-infra14:16
openstackgerritA change was merged to openstack-infra/config: Make python33 gate voting for solum project
*** yaguang has joined #openstack-infra14:16
*** roeyc has joined #openstack-infra14:17
*** yamahata_ has joined #openstack-infra14:18
mordredjeblair: ooh. ooh.14:18
mordredI just had a betterer though on the upload-wheels job14:18
mordredwe should write it to assume that the things it's going to make wheels of are universal wheels14:19
mordredand then only add the upload wheels jobs to our projects that are py2/py3 compliant14:19
mordredI believe that should cover most of our current concern with them14:19
mordredand make us not have to have quite so many confusing jobs14:20
*** ArxCruz has quit IRC14:21
sdaguewhere are ttx's release scripts for automatically doing things to lp artifacts?14:22
*** crank has quit IRC14:23
sdagueman, ttx doesn't keep it in gerrit :)14:26
*** rongze has quit IRC14:26
openstackgerritQianLin proposed a change to openstack-infra/nodepool: Fix the logprint in
mordredttx: maybe it's time to move that repo into gerrit14:30
mordredttx: even if it's not 'ready'14:31
*** rongze_ has joined #openstack-infra14:31
mordredttx: if I'm going to give bryce shit about the website, I should probably be consistent and give you shit at the same time14:31
* mordred officially gives ttx shit14:32
fungimordred: speaking of handing out complimentary fecal matter, are you also planning to be on the o'reilly press integration planning call today?14:34
mordredfungi: oh god. when is that?14:34
mordred1-1:30 eastern is it?14:34
*** xchu has quit IRC14:34
*** nicedice has joined #openstack-infra14:34
fungiyeah, 1800-1830z14:35
mordredon my calendar now14:35
mordredthank you14:35
*** roeyc has quit IRC14:35
mordredfungi: how far did you get with them last time?14:36
fungimordred: basically last time i had about 30 minutes to reexplain our workflow to them (because these aren't the same people we'd talked to originally there) and then figure out what could be done in a week while everyone else was out/busy since they had a deadline14:37
mordredok. so you've explained our workflow to them at least14:37
mordredwhich means today will be explain workflow #3 overall14:38
mordredand in another 2 months they might get it14:38
fungiyeah. i initially started to push them toward synchronous work, but they expressed major concerns about merge conflicts on docbook/xml content14:38
*** jergerber has joined #openstack-infra14:38
fungiand the complexity of trying to resolve those14:38
mordredwell, so here's the thing I think I'd like to bring home if we can ...14:39
*** jorisroovers has quit IRC14:39
mordredwhich is that we've done a ton of work to ensure that none of our contributors are more special than any of our other contributors14:39
ttxmordred: in my case it's officially because i'm overworked14:39
mordredsometimes to the pain of some of the contributors14:39
mordredfor that reason, it's very important that we do not have a special process for o'reilly because they are "special"14:39
ttxmordred: but yeah, I guess I can clean them up AFTER they are moved centrally14:39
fungittx: your new top priority is that cloning device which makes more ttxes14:39
ttxmordred: I always wanted to clean them up BEFORE moving them14:40
mordredttx: I can do the config patch to suck them in if you're fine with that14:40
mordredttx: I know the feeling14:40
ttxmordred: that said, they used to be in a bzr repo so it could be worse.14:40
*** sarob has joined #openstack-infra14:40
mordredttx: testr still is14:40
ttxmordred: I'm fine with you sucking my stuff in.14:41
fungimordred: actually a good chunk of the last call was trying to get them accustomed to the idea that "git" is not "github"14:41
ttxI might regret that one14:41
mordredfungi: so either a) we need to identify what's wrong with docbook/xml process for _all_ of our contributors and fix it, or b) o'reilly needs to use the process as is14:41
fungiexpect them to continue using github workflow terminology on this call14:41
mordredfungi: that is a thing I'm well equipped to handle14:41
mordredwe should start referring to the linux kernel workflow, which also does not use github14:41
fungiyes. we develop like the linux kernel, except we replace the mailing list with a code review system and replace the benevolent dictator with a distributed meritocracy of code reviewers14:42
*** yaguang has quit IRC14:42
*** mfer has joined #openstack-infra14:42
*** damnsmith is now known as dansmith14:43
fungithat's how i've started trying to explain it, though that's only helpful when explaining to people who are already familiar with linux kernel development (i.e. a very small cross section of the developer population)14:43
*** sarob has quit IRC14:45
*** roeyc__ has joined #openstack-infra14:45
mordredfungi: I'm going to tweet that as "We develop like Linux, except email is a code review system and the benevolent dictator is a meritocracy of code reviewers"14:46
mordred(140 character is hard)14:46
mordredor maybe14:46
mordred"We develop like Linux, except with a code review system and the benevolent dictator is a meritocracy of code reviewers"14:47
*** CaptTofu has quit IRC14:47
*** nsaje has quit IRC14:48
*** CaptTofu has joined #openstack-infra14:48
*** Shrews_ is now known as Shrews14:49
ttxmordred: I tend to think our benevolent dictator is a system centered around code review and strict gating14:49
ttxrather than a set of people14:49
*** rustlebee is now known as russellb14:49
ttxour dictator is a set of house rules that we collectively set14:49
markmcargument against that would be a robot can't design software :)14:49
markmci.e. that the benevolent dictator has a major design influence14:50
ttx(and that we accept to follow)14:50
*** dolphm has joined #openstack-infra14:50
ttxI'm looking into those models for a future presentation / blogpost14:51
ttxLinux is a BDFL-centered trust tree14:52
*** oubiwann has joined #openstack-infra14:52
ttxOpenStack is an organic set of development rules14:52
ttxgravitating around the gate14:52
ttxthat do not really involve people or trust as much14:52
markmcsure, that's the process side14:53
ttxwhich means it's more resilient to node failure14:53
markmcthe "OpenStack needs a BDFL" claims are more about leadership and design influence, though14:53
markmcnot that I agree with them :)14:53
markmcjust if you're arguing against those claims .... :)14:53
ttxmarkmc: yeah, it's strong design decisions they want14:53
ttxmarkmc: that said when we take them we also piss people off14:54
markmcttx, yeah, that's why I tend to shrug at this kind of stuff - listen, but *shrug*14:54
*** markmcclain has joined #openstack-infra14:54
markmcttx, more interested in improving our model than debating why another model is better/worse/analogous14:55
ttxmarkmc: I suspect that sort of strong decision goes better when it appears to be from some opiniated dude, rather than a consensus amongst level-headed people14:55
ttxmarkmc: because in the former there is little you can do14:55
*** markmcclain has quit IRC14:56
*** yaguang has joined #openstack-infra14:56
*** markmcclain has joined #openstack-infra14:56
mordredwell, people should look at decisions made by linus or sabdfl guido14:56
ttxwhile in the latter it looks like you could argue the set of people was wrong14:56
mordredand then re-think whether they think somethign can be done about devisions made by bdfl14:56
*** ArxCruz has joined #openstack-infra14:56
*** CaptTofu has quit IRC14:57
*** CaptTofu has joined #openstack-infra14:57
mordredI think also they think that some of our decisions are lack of decision, rather than explicit decisions to allow multiple viewpoints14:57
markmcit's rare to find someone sufficiently opinionated to be a bdfl who also understands how to foster a community14:57
fungiopenstack development is an emergent property of free software developers confronted with common/related challenges14:58
ttxanyway, "bdfl" was a great newspeak find. Replace it with plan "dictator" and you start seeing the drawbacks14:59
*** dims has joined #openstack-infra14:59
mordredttx: yah14:59
mordredttx: now _that_ is a good tweet15:00
ttxyou can be lucky (Linux, Guido) or unlucky (fill your favorite fail here)15:00
*** oubiwann has quit IRC15:00
mordredeven with guido - his insistence in not allowing the stackless patch15:00
mordredis what has us stuck with eventlet right now15:01
ttxmordred: true. It's really HARD to stay sufficiently on top of things to still embody the will of the other contributors15:01
mordredjeblair: and
ttxMost of those dictators slightly fade from view and then you start having a problem15:02
ttx(gentoo anyone)15:02
ttxBDFL are anomalies, historical artifacts. Not something you should plan to have in when building a new system15:03
ttxbasically, if you're asking yourself the question 'do we need one?', it's probably too late to have one anyway15:04
*** ryanpetrello has joined #openstack-infra15:07
*** beekneemech is now known as bnemec15:07
*** mrodden1 is now known as mrodden15:11
*** wenlock has joined #openstack-infra15:11
*** rcleere has joined #openstack-infra15:16
*** eharney has joined #openstack-infra15:18
openstackgerritMonty Taylor proposed a change to openstack-infra/config: Use requirements project list for mirror selection
openstackgerritDavid Peraza proposed a change to openstack/requirements: Adding sqlalchemy db2 dialect dependencies
*** datsun180b has joined #openstack-infra15:22
*** dcramer_ has joined #openstack-infra15:22
mordredyou know that they're LESS likely to have merge conflicts if they work syncrhonously right?15:23
mordredfungi: ^^ (that was me responding to a long time ago)15:23
*** crank has joined #openstack-infra15:24
fungimordred: yeah, i know. they seem to want to take a snapshot of the material, iterate on it within their group over the course of a week or two, then dump it back to us once they're happy with it. not particularly compatible with how we work15:25
mordredfungi: what do they expect our material to do in the meantime?15:26
fungimordred: i think they are used to getting work in its near-finalized state from the author, ready for editing prior to mass publication15:27
*** oubiwann has joined #openstack-infra15:29
mordredso, it sounds similar to how we do stable branches15:29
mordredlike, cut stable branch, then a new/smaller set of people get +215:29
*** markmcclain has quit IRC15:30
*** xeyed4good has joined #openstack-infra15:30
fungiright, except given that their edits are going to be, for the most par, corrections... we'd like to be able to capture those corrections and feed them back into the development version15:31
*** boris-42 has quit IRC15:31
fungiso there'd still be cherry-picking of that back into master15:31
fungiso maybe more akin to a feature branch than a stable branch15:31
*** nsaje has joined #openstack-infra15:34
*** thedodd has joined #openstack-infra15:34
*** markmcclain has joined #openstack-infra15:37
*** sarob has joined #openstack-infra15:40
*** dhellmann is now known as dhellmann-afk15:41
*** ryanpetrello has quit IRC15:42
*** markwash has quit IRC15:44
*** sarob has quit IRC15:44
openstackgerritAndreas Jaeger proposed a change to openstack-infra/config: Readd object-storage-manual for Grizzly
*** CaptTofu has quit IRC15:47
*** yaguang has quit IRC15:47
*** afazekas has quit IRC15:48
ttxhere we go again:
*** hashar is now known as hasharAW15:49
openstackgerritA change was merged to openstack-infra/reviewstats: Tweak formatting for hacking.
openstackgerritA change was merged to openstack-infra/reviewstats: Cookiecutterify reviewstats.
*** atiwari has joined #openstack-infra15:52
*** UtahDave has joined #openstack-infra15:52
openstackgerritA change was merged to openstack-infra/reviewstats: Ignore E126.
openstackgerritA change was merged to openstack-infra/reviewstats: Move tests into the setuptools structure.
openstackgerritA change was merged to openstack-infra/reviewstats: Move utils into the reviewstats namespace.
*** dkranz has quit IRC15:54
*** cyeoh has quit IRC15:56
*** dolphm has quit IRC15:57
mordredttx: I vote for jekyll - mainly because jekyll island is pretty awesome15:57
*** julim has quit IRC15:57
mordred(I've been there many times - my cousin lives in brunswick and got married on jekyll)15:57
ttxYeah, jekyll is a strong contender for so many reasons15:57
mordredttx: you know, we could do the k vote too15:57
ttxlooks cool, sounds cool, is cool and has a personality disorder15:57
ttxIf nto Jekyll, I would pick "Jersey"15:58
mordredyah - but that would be the second name tat sounded like a name of another famous place15:58
mordredwhich maybe is fun15:59
ttxI kinda like to be confusing15:59
ttxFor K we could pick Kourou, which is a French city but in South America. Which is confusing too.15:59
ttxAlso home to the European Space Center where they launch Arianes and Soyuzes15:59
mordredttx: we could also go with Julep - which is a name of a popular alcoholic drink in Georgia15:59
*** julim has joined #openstack-infra16:00
ttxthen we should be out of the poor zone, until we hit W/X/Y/Z16:00
mordredttx: also I got the suggestion that we go with "junior" to pay homage to Martin Luther King Junior16:01
ttxI still want us to stick with the rules, so unless Julep or Junior is on the state flag, no.16:01
*** danger_f` is now known as dangers16:01
ttxwithout rules, naming is not fun16:01
*** hasharAW is now known as hashar16:03
mordredttx: there is a "John F Kennedy Middle School" in atlanta. that's a place ... could we call the release "John F Kennedy" ?16:03
* mordred ducks16:03
*** jgrimm has joined #openstack-infra16:03
mordredwe could also pick juniper for juniper street  - and wait to see how the juniper networks guys react :)16:04
mordredI feel like we kinda have to see what markmcclain thinks too16:05
ttxmarkmcclain, I think was a pretty strong Jekyll fan16:05
markmcclainI live a block from Juniper16:06
ttxyes, but no.16:06
markmcclainwould be fun to see the lawyer response to that one16:06
ttxIf I didn't know the answer I could probably find it funny too16:06
*** cyeoh has joined #openstack-infra16:06
markmcclainttx: the MLK Jr Blvd … so wouldn't junior be allowed?16:07
ttxmarkmcclain: we use street names ONLY if there aren't any suitable city/county names.16:07
*** dkranz has joined #openstack-infra16:08
mordredwell then, I think jekyll is the best currently posed choice16:08
markmcclainJasper was my 2nd option16:08
mordredyeah. I could go with jasper16:08
ttxalthough jaspersoft would become a concern16:09
openstackgerritSean Dague proposed a change to openstack-infra/elastic-recheck: added bug #1210483 detection
uvirtbotLaunchpad bug 1210483 in nova "ServerAddressesTestXML.test_list_server_addresses FAIL" [Undecided,Fix released]
*** DinaBelova has joined #openstack-infra16:10
mordredsdague: have we seen this one much:16:11
mordredSSHTimeout: Connection to the via SSH timed out.16:11
mordredUser: cirros, Password: None16:11
sdaguedo you have a more specific instance of the fail?16:11
markmcclainI noticed that popped up on a translation patch16:11
mordredpopped up on a pbr patch16:11
sdaguemordred: so we just pushed a change that turned back on the ssh enabled flag for devstack/tempest16:12
mordredsdague: AWESOME16:12
mordredI thnk that's super important16:12
mordredso I'm guessing we might find a few more races16:12
*** mrodden has quit IRC16:12
sdaguemordred: ok, but it would be good to figure out if that issue is seen pre this morning in any quantity16:13
mordredwell, it's not possible for it to have been is it? if that's a ssh-to-image issue16:13
*** pblaho has quit IRC16:13
luisgi have several patches Approved, gate started nov 20, but they are still not merged, they have been like that for a while, is there anything i need to do to get them merged?16:13
mordredsdague: logstash only shows one entry for SSHTimeout16:14
*** yolanda has quit IRC16:14
sdaguemordred: you set the time slider right?16:15
sdagueit defaults to 15 mins16:15
*** senk has joined #openstack-infra16:15
markmcclainshowing 14 for last 7d16:16
sdagueok, let's get a bug registered, and an er fingerprint16:17
sdaguemy goal is 90% coverage in er of all resets16:17
openstackgerritRussell Bryant proposed a change to openstack-infra/reviewstats: Basic usage instructions.
openstackgerritRussell Bryant proposed a change to openstack-infra/reviewstats: Move reviews_for_bugs into reviewstats.
openstackgerritRussell Bryant proposed a change to openstack-infra/reviewstats: Move openapproved into the reviewstats namespace.
openstackgerritRussell Bryant proposed a change to openstack-infra/reviewstats: Move openreviews into the reviewstats namespace.
openstackgerritRussell Bryant proposed a change to openstack-infra/reviewstats: Move into the reviewstats namespace.
openstackgerritA change was merged to openstack-infra/reviewstats: Basic usage instructions.
*** adam_g_afk is now known as adam_g16:19
*** adam_g has quit IRC16:19
*** adam_g has joined #openstack-infra16:19
pabelangerQ: What is the preferred way to ignore a directory in .gitignore, if the directory is foo.  I remember hearing .foo/* in the past, however an open review has other opinions about it16:21
sdagueis er bot CD?16:22
*** che-arne has quit IRC16:23
openstackgerritA change was merged to openstack-infra/elastic-recheck: added bug #1210483 detection
uvirtbotLaunchpad bug 1210483 in nova "ServerAddressesTestXML.test_list_server_addresses FAIL" [Undecided,Fix released]
*** mrodden has joined #openstack-infra16:25
*** mihgen has quit IRC16:25
fungiluisg: we had to drop the gate queue on the 20th to deal with a bunch of nondeterministic errors in various parts of openstack which were gumming up the works. it's safe to just leave a review comment on each of those which says only "reverify no bug" and they should get added back into the gate immediately16:25
fungisdague: i believe so, except that it may not restart automatically on code changes only config changes16:26
fungipabelanger: foo/16:28
fungipabelanger: er, no. foo (without the /) should be enough16:29
fungithough apparently both forms work as of 1.5.516:29
*** DinaBelova has quit IRC16:30
*** dcramer_ has quit IRC16:31
*** pblaho has joined #openstack-infra16:31
*** pblaho has quit IRC16:32
*** boris-42 has joined #openstack-infra16:32
*** che-arne has joined #openstack-infra16:32
*** davidhadas has quit IRC16:32
pabelangerfungi, I remember the thinking for foo/* was to indicate a directory to the developer, vs foo directly16:33
*** wenlock_ has joined #openstack-infra16:33
*** wenlock_ has quit IRC16:33
*** wenlock_ has joined #openstack-infra16:33
luisgfungi, i see thank you16:34
openstackgerritPaul Belanger proposed a change to openstack-infra/zuul: Ignore .venv directory
fungipabelanger: foo/* will only ignore the contents of foo but will not ignore foo itself, i believe16:34
pabelangerfungi, right, but I don't think git will allow you to add an empty directory16:35
pabelangerI've updated my review for .foo however, thanks for the feedback :)16:35
fungigood point!16:35
fungii think any of them is probably acceptable, personally16:35
*** sarob has joined #openstack-infra16:40
openstackgerritPaul Belanger proposed a change to openstack-infra/zuul: Fix H hacking violations
*** pblaho has joined #openstack-infra16:43
openstackgerritPaul Belanger proposed a change to openstack-dev/cookiecutter: Remove install_venv_common module
*** wenlock__ has joined #openstack-infra16:47
jeblairpabelanger: I don't want to enforce hacking on infra projects.16:47
*** dcramer_ has joined #openstack-infra16:47
pabelangerjeblair, Oh, I did not know that16:47
*** SergeyLukjanov is now known as _SergeyLukjanov16:47
jeblairpabelanger: yeah, sorry.  at the summit we decided that we need to put comments in all of the tox.ini files saying if we are ignoring something on purpose, or because we just haven't gotten around to cleaning it.16:48
*** _SergeyLukjanov has quit IRC16:48
jeblairpabelanger: had we done so, i suppose the 'H' ignore should be commented thusly.16:48
*** sandywalsh has quit IRC16:49
pabelangerjeblair, fair enough. Was about to cookiecutterify zuul, should I bother?16:49
*** wenlock_ has quit IRC16:49
zulim seeing some pbr weirdness
jeblairpabelanger: anyway, i think the change itself looks fine, except for the ignore line in tox.ini; i'll leave a comment to that effect16:49
*** blamar has joined #openstack-infra16:49
jeblairpabelanger: what does it mean to cookiecutterify something?16:49
*** guohliu has joined #openstack-infra16:50
pabelangerjeblair, formatting of zuul based on what openstack-dev/cookiecutter generates. Again, an attempt to clean it up per other openstack project.  But if you don't see the value, I'll move on16:51
jeblairpabelanger: what kind of formatting?  i thought cookiecutter just generated some stub files.16:51
pabelangerjeblair, Right, stub files, documentation, more dependency on oslo-incubator.git.16:52
jeblairpabelanger: i'm having a hard time imagining how depending on oslo-incubator would be beneficial for us, or what kinds of stub files or documentation zuul is lacking; i'd suggest...16:54
jeblairpabelanger: if you want to look into that, please do so with a critical eye.  i would not assume that everything it does is appropriate.  i'm particularly skeptical that tying to oslo-inc would be productive16:55
*** senk has quit IRC16:55
pabelangerjeblair, that is fine. Like I said, I see some other -infra projects doing clean up recently ( figured I do the same to zuul.16:56
pabelangerIf there is no value, then no need to do it.16:56
*** ruhe has quit IRC16:56
*** nsaje has quit IRC16:56
*** nsaje has joined #openstack-infra16:57
jeblairpabelanger: i haven't looked at cookiecutter output recently, so i can't fully evaluate the value; but i will say that reviewstats is a newish project with very little infrastructure and docs, which isn't the case for zuul and jjb.  nodepool might benefit.16:59
pabelangerjeblair, okay, maybe what I'll do is bring it up in the next -infra meetings. Basically, looking to get more involved here and was just working on low hanging fruit.17:00
*** pblaho has quit IRC17:01
*** jergerber has quit IRC17:01
*** zoresvit has quit IRC17:01
*** zoresvit has joined #openstack-infra17:02
jeblairpabelanger: thank you.  i'm almost certain nodepool could benefit from cookiecutter.17:02
*** nsaje has quit IRC17:02
mordredsdague: oops. nope. I did not timeslider it :)17:02
*** sandywalsh has joined #openstack-infra17:02
jeblairpabelanger: also, i can't overstate the value of your puppet reviews.  :)17:03
ArxCruzjeblair: hey, can you take a look at this:
jeblairArxCruz: i think i can't honor requests to look at reviews now17:03
jeblairunless it's an emergency17:03
ArxCruzjeblair: oh :(17:03
jeblairi have a huge backlog, and i think part of it is because people treat reviews like a stack, and not a queue17:04
*** jorisroovers has joined #openstack-infra17:04
ArxCruzjeblair: sorry :-#17:04
*** SergeyLukjanov has joined #openstack-infra17:04
mordredpabelanger: puppet reviews ++17:05
* pabelanger nods17:05
*** jroovers has joined #openstack-infra17:06
*** dkranz has quit IRC17:07
*** dkranz has joined #openstack-infra17:07
*** jorisroovers has quit IRC17:08
*** dkranz has quit IRC17:11
*** dkranz has joined #openstack-infra17:11
*** dolphm has joined #openstack-infra17:12
jeblairannegentle, ttx: why is there a 'nearly/havana' branch of openstack-manuals?17:13
*** sarob has quit IRC17:13
jeblairpleia2: regarding did you test that devstack worked after preparing a centos node with that change?17:17
*** davidhadas has joined #openstack-infra17:18
*** wenlock__ has quit IRC17:18
*** wenlock_ has joined #openstack-infra17:18
mordredoh wow17:20
mordredmy pbr change got results from redhatci17:20
mordreddprince: do you know those guys ^^ ?17:21
*** Ryan_Lane has joined #openstack-infra17:21
dprincedprince: Yes. I think its Ian.17:21
*** fbo is now known as fbo_away17:22
mordreddprince: well, I think it's awesome - and would be awesomer if the log links were served in such a way that I didn't have to download them :)17:22
*** wenlock_ has quit IRC17:23
zaroclarkb: just read your post-mortem. that was really good.17:23
*** reed has joined #openstack-infra17:23
*** dkranz has quit IRC17:24
*** jroovers has quit IRC17:24
clarkbmordred: we are not universal py2 and py3 fwiw (we skip several releases in the middle)17:24
dprincemordred: I passed along the info. Not sure Ian is online right now...17:24
clarkbfungi: sdague: er is CD for both config and code17:24
clarkbzaro: jog0 did the hard work of making it readable.17:25
fungiclarkb: it successfully restarts on updates of either now?17:25
clarkbfungi: yes (config changes don't need a restart but code updates will trigger a service restart)17:25
clarkbmordred: why aren't we talking about 'Jenkins' more seriously :)17:26
fungijeblair: the nearly/havana branch was being used by the docs team just prior to release to exercise documentation macros which extracted the release name from the branch (they didn't want an early stable/havana branch for fear of confusion), if memory serves. i noticed it again when i was deleting the stable/folsom branches but didn't delete it because i wanted to confirm with annegentle that she no17:28
fungilonger needs it17:28
*** davidhadas has quit IRC17:28
*** pete5 has joined #openstack-infra17:30
openstackgerritPeter Liljenberg proposed a change to openstack-infra/jenkins-job-builder: Added support for Jenkins plugin Blame upstream committers
openstackgerritArx Cruz proposed a change to openstack-infra/devstack-gate: Fedora/RHEL DB updates for
*** markmcclain has quit IRC17:31
*** markmcclain has joined #openstack-infra17:31
jeblairclarkb: dprince change at touches devstack jobs17:32
zaroclarkb: dcramer_ says he's ready for us to clone maven clouddocs plugin.  you want to do that?17:32
mordredclarkb: I think I'm fine with the level of universalness that we are - but I'm also fine backing off of that and just never doing them17:32
clarkbjeblair: looking, thanks17:32
openstackgerritA change was merged to openstack-infra/config: remove code duplication in run-*.sh
clarkbmordred: it may not be worth worrying about considernig the install bases of 3.1 and 3.217:33
mordredclarkb: that's kinda what I was thinking about17:33
clarkbzaro: I probably won't get to that today, but can probably poke at it later this week17:33
mordredclarkb: a single universal wheel for 3.1 and 3.2 won't be any worse than a source tar.gz :)17:33
*** markmcclain has quit IRC17:33
openstackgerritPeter Liljenberg proposed a change to openstack-infra/jenkins-job-builder: Cleaned up fixtures. Removing vim editor settings
*** swifterdarrell has left #openstack-infra17:34
zaroclarkb: ok, no hurry.  if someone else can do that would work as well.17:34
*** roeyc__ has quit IRC17:35
lifelessrussellb: thanks for fixing up the usage docs17:35
*** hashar has quit IRC17:35
clarkbzaro: maybe bring it up in tomorrow's meeting if it hasn't been done yet. I really need to finish the d-g job refactor so that jeblair doesn't need to -1 all of those patches :)17:35
*** CaptTofu has joined #openstack-infra17:36
*** gyee has joined #openstack-infra17:36
*** dkranz has joined #openstack-infra17:36
*** jpich has quit IRC17:36
*** dcramer_ has quit IRC17:36
mordredclarkb, zaro: what's the task?17:36
lifelessrussellb: though, since the way most folk will run it is the shell scripts, I'm not sure we shouldn't document the variable ;)17:37
russellblifeless: np, was going to work through the rest of them today too17:37
*** dolphm has quit IRC17:37
russellblifeless: i figured i was the only one using the shell scripts17:37
russellband everyone else runs them one-off to check in on their project of choice17:38
lifelessrussellb: I have no idea :)17:38
lifelessrussellb: anyhow17:38
clarkbmordred: updating clouddocs maven plugin in gerrit17:38
russellbactually, i think most people look at the already generated stuff and don't bother17:38
russellbbut yes, thanks a bunch for this patch series, great stuff.17:38
mordredclarkb: as in, clone locally, force push to gerrit?17:38
clarkbmordred: and coordinating the actual stoppage of work in github17:38
clarkbmordred: yes, btu I think we will also need to do a reasonable job of firing off emails to the docs list and asking david to make the github repo read only otehrwise what happened before will happen again17:39
*** olaph has joined #openstack-infra17:39
mordredzaro: if dcramer is ready to stop work in github, I can do the clone/force-push17:39
lifelessrussellb: I'm thinking of rearranging the internals to suck down all the changes and then filter locally - build all the reports in parallel.17:39
lifelessrussellb: I wanted a stable base to do that on, which is what led me to the series :)17:39
lifelessrussellb: what do you think ?17:39
*** sarob has joined #openstack-infra17:40
russellbyeah makes sense.  there's already some code that caches the entire history on disk17:40
russellbbut not used for everything17:40
russellband i also wonder if we could just request all updates since what we have locally ... haven't looked at that yet17:41
russellbdownloading all of nova for example takes a while ...17:41
lifelessso where I'm really going with this17:41
lifelessis that to do the bug stats I want17:41
russellb-rw-rw-r--.  1 rbryant rbryant  50M Nov 21 17:11 .nova-changes.pickle17:42
lifelessI either need to implement more of my long pending causes project (fun, but incidental), or I need something to pull down a significant chunk of LP bugs17:42
lifelessand since there is some bug stats in here already17:42
mordredzaro: if you can coordinate with the dcramer for the work stoppage on github, I;m prepared to do the push whenever17:42
*** sarob has quit IRC17:44
*** ^d has joined #openstack-infra17:48
*** dcramer_ has joined #openstack-infra17:48
*** ^d has joined #openstack-infra17:49
zaromordred: dcramer's email:
mordredzaro: ok. I'm going to force-push up17:51
mordredzaro: remind him that it'll be stackforge/clouddocs-maven-plugin - not openstack/17:51
openstackgerritJames E. Blair proposed a change to openstack-infra/config: Document what it takes to be a check/gate test.
clarkbjeblair: fungi: mordred: so the current problem with the d-g job reorg is figuring out how to collapse the check jobs for stable branches into the templates for the normal jobs. The check jobs for stable branches override ZUUL_BRANCH17:51
zaromordred: yup. already did that. thanks..17:51
*** afazekas has joined #openstack-infra17:51
fungijust a heads up, i've -2'd for now, until bug 1228977 is addressed17:51
uvirtbotLaunchpad bug 1228977 in nova "n-cpu seems to crash when running with libvirt 1.0.6 from ubuntu cloud archive" [High,Confirmed]
*** yolanda has joined #openstack-infra17:51
jeblairfungi: thank you17:52
fungiit had a pair of +2s, so i wanted to make sure it doesn't inadvertently get merged17:52
lifelessrussellb: oh, I also registered reviewstats on pypi and added full check gate and release jobs to -config17:52
clarkbjeblair: fungi: mordred: we could continue to treat those jobs as different, or I could add logic to the shell bits run by d-g jobs that checked flags set by JJB17:52
clarkbI don't really care for either option and am open to other suggestions if folks have them17:53
lifelessrussellb: (and added openstackci as owner, of course)17:53
fungiclarkb: i'm not coming up with any cleaner alternatives at the moment, but will admit i haven't considered it a ton yet. i'll ponder17:53
mordredzaro: done17:53
mordredzaro: is the review to add .gitreview17:53
jeblairclarkb: jjb setting ZUUL_BRANCH is just convenience; if you want to change it to something more semantic, i think that would be fine.17:53
*** derekh has quit IRC17:54
russellblifeless: aewsome17:54
russellbit's growing up to a real project, *tear*17:54
*** ryanpetrello has joined #openstack-infra17:55
jeblairmordred: did you get fake testing on working?17:56
*** markwash has joined #openstack-infra17:56
zaromordred: i created a patch before the re-clone, do you think it will work? or do i need to abandon and create new one?17:57
jeblairzaro: you can hit the 'rebase' button and find out17:57
*** melwitt has joined #openstack-infra17:57
zaroahh, it failed.  i will do the manual rebase.17:58
clarkbjeblair: fungi: is what I am currently looking at17:58
mordredrussellb: god. next thing you know I'll actually get around to running it in infra for you ...17:58
russellbthat would rock17:58
clarkbjeblair: fungi: set the JJB var to 'default' to get normal behavior or change it to an actual branch name if you want to override the zuul branch17:58
russellbi know i could set that up too .... i just haven't17:58
russellbhard to get motivated when it's working for now17:58
mordredjeblair: no. when you have time, I'd love to take a few minutes with you to figure that out and understand why what I was doing wasn't working17:59
mordredjeblair: the code for doing it looked very straightforward17:59
mordredso I was sad when it did not do the thing I expected it to do17:59
jeblairmordred: ok.  let's check that out after the call, which is about to start17:59
mordredjeblair: I'm so excited about this call, I think I'm going to poo17:59
*** nsaje has joined #openstack-infra18:00
jeblairit sounds like the asterisk lady!18:01
*** wenlock_ has joined #openstack-infra18:04
*** sarob has joined #openstack-infra18:04
harlowjaanyone mind giving a 'approve' to :)18:06
*** salv-orlando has quit IRC18:10
*** jorisroovers has joined #openstack-infra18:11
*** osanchez has quit IRC18:12
*** jroovers has joined #openstack-infra18:14
*** thomasem has quit IRC18:14
openstackgerritKhai Do proposed a change to openstack-infra/config: new maven plugin jobs to build and deploy maven plugins to maven nexus
*** jorisroovers has quit IRC18:16
*** mihgen has joined #openstack-infra18:16
ArxCruzdprince: around ?18:17
ArxCruzdprince: can you take a look at jeblair is waiting a response from you :)18:18
openstackgerritKhai Do proposed a change to openstack-infra/config: new maven plugin jobs to build and deploy maven plugins to maven nexus
jeblairArxCruz: or anyone who wants to volunteer to spin up a precise node and test :)18:18
blamaranyone know how much RAM a machine has to have to run Neutron tests? I just had tests chew through 2GB of RAM and 2GB of swap to run out of memory...18:19
ArxCruzjeblair: my tests are on power, I only have fedora 1918:19
ArxCruzjeblair: besides, if is approved, I can remove both yum and apt-get for that patch and it will work :)18:19
clarkbblamar: I think the answer is lots, and unfortunately it has been that way for a long time. Our test slaves have 8GB (they are that large so that we can have 4 cpu per slave not because we thinkit should take that much memory to run unittests)18:20
clarkbblamar: we might want to look at stronger memory constraint enforcements with cgroups18:20
clarkbwe currently keep tests to a hard limit of 95% of RAM to help protect slaves from unfortunate OOMing18:20
*** xeyed4good has quit IRC18:20
dprinceArxCruz: Um. Okay. I didn't realize that was specifically to me. I suppose I *could* do it, but I don't have a precise node handy (nor do I usually ever)18:21
pleia2jeblair: re:54386 I'll give centos a spin in devstack specifically (figured it was essentially the same command as in ubuntu so only tested the command on an fresh centos hpcloud instance, not w/ devstack specifically)18:22
ArxCruzdprince: jeblair sorry, my mistake it's
blamarclarkb: thanks for the info -- was really hoping to get by with 2GB slaves but I guess I'm just kidding myself18:22
ArxCruzdprince: lot of ctrl+c ctrl+v here :)18:22
ArxCruzdprince: I need to leave now, I will be back in a few hours, I would appreciate if you take a look in
*** ArxCruz has quit IRC18:26
*** dizquierdo has quit IRC18:29
*** nsaje has quit IRC18:31
*** thomasem has joined #openstack-infra18:31
annegentleOMG WHAT HAVE I DONE? :)18:31
annegentleJust kidding18:31
zaroany core infra available to help with this?
zarowe need to deploy to maven nexus, so will need an account for that.18:33
openstackgerritA change was merged to openstack-infra/config: Remove extra chars in launchpad bug links
annegentlejeblair: so merging in the whole branch sounds like an okay solution but a bit of a leap of faith "we'll get what we get" I guess18:33
*** markmc has quit IRC18:33
jeblairannegentle: either it will just work, or it will be a big mess you have to clean up, or something in the middle.. but...18:33
annegentlejeblair: yeah. :)18:34
annegentlejeblair: worth trying though? and but what? :)18:34
mordredzaro: so thigns needed are: a) infra-core to make a maven nexus account and put the info into heira18:34
*** markmcclain has joined #openstack-infra18:34
mordredzaro: then landing that change, yeah?18:34
jeblairannegentle: if they keep their branch up to date as they work, you can continually try to merge it into master18:34
zaromordred: yup.18:34
mordredzaro: how urgent is that? we've got a little bit of time, yeah?18:34
annegentlejeblair: that's what I hope, they sound amenable to that18:34
jeblairannegentle: and you'll get a handle on how significant a problem it is (and basically have the ability to break any work that there is up into sizable chunks of your own choosing)18:35
annegentlejeblair: yes that is ideal, having the ability to size the merges. YES.18:35
zaromordred: no hurry.  actually i think that patch might need account in heira for it to work?  we can deal with account on maven nexus only if we want to release.18:36
zaromordred: site.pp references hiera accounts that don't exist.  not sure if it will fail or just return empty.18:37
openstackgerritA change was merged to openstack-infra/config: Document what it takes to be a check/gate test.
*** afazekas has quit IRC18:39
clarkbjeblair: is what you had in mind for making it more semantic? Now that I have written it out I don't find it too bad18:39
jeblairmordred: provider_manager iterates over a custom dict it makes of the flavors, not the actual flavor objects returned18:39
*** wenlock_ has quit IRC18:39
jeblairmordred: see _getFlavors()18:39
*** wenlock_ has joined #openstack-infra18:40
jeblairmordred: it was probably to avoid keeping a bunch of novaclient objects around forever; i don't recall if there was a stronger reason for that18:40
clarkband the periodic jobs can be made to be very similar with a few differences (because they are periodic)18:40
*** rongze_ has quit IRC18:40
jeblairmordred: anyway, i think once you see _getFlavors you'll grok it.18:41
*** CaptTofu has quit IRC18:41
openstackgerritA change was merged to openstack-dev/pbr: Serve local mirror using apache
jog0fungi: ping18:42
clarkbjeblair: fungi: if that paste doesn't look too horrible to you guys I will go ahead and do a mass conversion of the jobs, as I think that will get me to where I think the jobs should be18:42
jeblairclarkb: lgtm; is it worth thinking about old and new branch overrides for grenade now, or baby steps?18:42
jeblairclarkb: ++18:42
clarkbjeblair: baby steps :)18:42
jog0fungi: for your email about libvirt 0.9.1118:43
jog0'. I'm consistently getting18:43
jog0numerous tempest tests failing with "server in error state" messages18:43
jog0on the console'18:43
jog0any more detail on that18:43
clarkbjeblair: I have a hunch the grenade jobs will need to be a third set of thing with a bunch of different considerations, easier to ignore that for now to avoid confusion18:43
sdaguehow much I want to punch launchpad for timing out every 6th operation18:43
fungijog0: i pasted the traceback into the bug, but it's trivial to recreate18:44
*** jcoufal has quit IRC18:44
jog0fungi: ahh that log is useless, any nova logs?18:45
jog0fungi: either way that answers my question. I wasn't sure what you ment about console, I thought that it may be the console log like the *920 bug18:45
jog0but its not, thanks18:45
fungijog0: i have destroyed the virtual machines on which i ran those tests, but it happened 100% of the time. i just wasn't sure what specifically to collect from them or where to look for the underlying cause18:47
jog0fungi: no problem, in the future all of nova logs18:47
jog0is a good place to start18:47
jog0anyway just wanted to make sure it was an unrelated issue18:48
fungijog0: i was mainly just trying to confirm whether or not we can turn ubuntu cloud archive back on, and it seems we can't. my guess is it's probably something with new python-libvirt (since that was the blocker the last time we tried) but i can't be sure18:48
pete5I had a gate test fail because of a timeout creating an instance. The instance was successfully created, but it took 1 minute longer than tempest was willing to wait. Can I do a "reverify no bug" for this, or should I file a tempest bug saying that the timeout is too short? Here's the failure: and here's the log showing th18:48
pete5it was ready 1 minute later:
openstackgerritSean Dague proposed a change to openstack-infra/elastic-recheck: add fingerprint for bug #1250168
jog0pete5: neutron-large-ops isn't voting yet18:49
uvirtbotLaunchpad bug 1250168 in neutron "gate-tempest-devstack-vm-neutron-large-ops is failing" [Critical,Confirmed]
fungithat was a funny coincidence18:49
pete5jog0: that's what i thought, but gerrit says otherwise:
jog0jeblair: can you take a look at'18:49
jog0pete5: AFAIK that is a bug18:50
jog0unless we made large-ops gating recently18:50
jog0neturon lagre ops18:50
fungijog0: it is voting as of friday i think (it was passing consistently, so i approved the change)18:50
jog0fungi: oh cool, works for me18:50
jog0didn't know thats all18:50
jeblairjog0: i'm only taking emergency requests to look at reviews out of order, sorry18:50
jog0jeblair: ack, this is not an emergency so never mind18:51
jeblairjog0: though you may be interested in the one i'm looking at right now: references bug
uvirtbotLaunchpad bug 1243251 in ceilometer "QueuePool ERROR in ceilometer-collector  log after successful tempest run" [Critical,In progress]18:51
jog0(it does need a review from you because the graphite js lib -- but not an emergency)18:52
jog0pete5: so looks like sdague just filed a bug18:52
*** nsaje has joined #openstack-infra18:52
jog0but *don't* do reverify no bug18:52
fungijog0: yeah, that was the amusing coincidence i was referencing (gerritbot announced it just as you said it was probably a valid bug)18:53
reedRyan_Lane, can you give me more details about the installation of Translatewiki on our wiki?18:53
sdaguejog0: it's your bug -
uvirtbotLaunchpad bug 1250168 in neutron "gate-tempest-devstack-vm-neutron-large-ops is failing" [Critical,Confirmed]18:53
Ryan_Lanelet me make sure it's actually active :)18:53
reedRyan_Lane, cool, let me know... It would be good to follow up on the i18n mailing list with it18:54
sdaguethere is apparently debate on it though, but until someone points me at a better bug, that's the one I'm going with18:54
Ryan_Laneyeah, it's active18:54
jog0sdague: that bug is apretty generic catch all bug18:54
Ryan_Lanereed: here's the details:
jog0but yeah18:55
sdaguejog0: sure, but it seems to be the only one that's being worked18:55
reedRyan_Lane, how do I learn how to translate pages?18:55
openstackgerritMonty Taylor proposed a change to openstack-infra/nodepool: Fix old pre-nodepoold reference in README
openstackgerritMonty Taylor proposed a change to openstack-infra/nodepool: Add ability to filter on flavor name
mordredjeblair: makes sense - I think the patch above now addresses that18:55
mordredjeblair: and I _think_ I have appropriately tested that it works18:55
mordredjeblair: although I would love validation of that18:56
jog0sdague: yeah I agree, sounds like a reasonable bug to fingerprint on18:56
jog0until we get more insight18:56
Ryan_Lanereed: I'm asking the translatewiki folks if they can give you a quick tutorial18:57
reedRyan_Lane, cool, thanks18:57
reedif they have a wiki channel I can join it18:57
*** sarob has quit IRC18:57
pete5jog0: thanks!18:57
*** sarob has joined #openstack-infra18:58
*** chandankumar has quit IRC18:58
jog0pete5: the main behavior we want to avoid, is oh gate failed, let me just reverify without seeing if the bug is my fault or not18:58
jeblairmordred: i just reviewed you may want to WIP or -2 it or something if you don't want it merged.18:59
pete5jog0: I 100% agree with that! So I spent about 15 minutes looking through the logs before messaging you guys :-)18:59
jog0pete5: :)18:59
sdaguejog0: we need to come up with a shorter name for gate-tempest-devstack-vm-neutron-large-ops19:01
sdaguebecause it's breaking the zuul ui19:01
jog0sdague: heh,  gtdvnlo19:02
jeblairi'd suggest dropping 'devstack-vm' but we're on the verge of finally being able to test something not in a devstack-vm.  but nonetheless, maybe we can make that shorter, with just 'ds' or something.19:02
jeblairalso, 'lops' sounds fun.19:03
clarkbjeblair: sdague: my d-g refactor is making the names even longer >_> but we can sort that out once we have agreement on the technical bits19:04
jeblairannegentle: are you still using 'nearly/havana'?19:06
*** markmcclain has quit IRC19:09
*** herndon has quit IRC19:09
fungijeblair: clarkb: mordred: it looks like we've somehow split nova, ceilometerclient and swift into a separate queue from the main integrated one. did we mess up on job renames?19:09
mordredfungi: that does not seem like a good thing19:10
fungior forget to add some jobs to the fake z/tempest project maybe?19:10
fungii'm starting to look, but more eyes would help19:10
mordredthe jobs look right to me?19:11
sdagueso I have a style work around that will at least make it less confusing, still not the prettiest, let me push19:11
*** rongze has joined #openstack-infra19:11
jeblairfungi: 'gate-tempest-devstack-vm-full' should cause them to be together...19:12
jeblairfungi: (i say that to indicate this is more likely to be a bug than a config file error)19:13
fungijeblair: agreed. so something going on zuul-side seems more likely19:13
openstackgerritSean Dague proposed a change to openstack-infra/config: create a jobwrapper span to address long job names spillover
*** rongze has quit IRC19:17
openstackgerritJoe Gordon proposed a change to openstack/requirements: Unpin keyring, just skip keyring 2.0
sdaguefyi - the top item is what it looks like after the wrap class in my above patch is added, the one below it is how it looks now19:19
*** nsaje has quit IRC19:20
fungilooks like zuul last reloaded its config about 5 hours ago (14:38:50z)19:20
*** nsaje has joined #openstack-infra19:21
*** odyi has joined #openstack-infra19:21
jeblairi'm debugging locally19:21
openstackgerritDan Prince proposed a change to openstack-infra/config: Update to install RPMs.
openstackgerritDan Prince proposed a change to openstack-infra/config: Add Fedora 19 to the nodepool (devstack-f19-check)
*** boris-42 has quit IRC19:22
*** DinaBelova has joined #openstack-infra19:22
*** nsaje_ has joined #openstack-infra19:24
jeblairfungi: i can't reproduce locally :(19:24
*** nsaje has quit IRC19:24
fungii'm not seeing anything in the logs which jumps out at me19:24
jeblairfungi: i'm going to send it the reload signal and let's see if it behaves differently19:24
jeblairfungi: it merged them19:26
mordredjeblair: _excellent_19:26
mordredjeblair: I think we should blame the GIL19:26
jeblairmordred: better than any idea i can come up with19:27
mordredjeblair: ++19:27
*** nsaje_ has quit IRC19:29
*** sarob_ has joined #openstack-infra19:29
jeblairfungi: looking at the logs under "Shared change queues:" the result it came up with this time looks sane19:30
fungijeblair: it definitely screwed it up on the previous reload, though i'm not (yet) finding why...
*** senk has joined #openstack-infra19:30
*** sarob has quit IRC19:30
sdaguemordred: did you register a bug for the SSH fail?19:31
* sdague realizes today is just - categorize all the uncategorized gate fails he sees19:31
*** sarob has joined #openstack-infra19:31
jeblairfungi: i wonder if loading n-1 layout and then n would exhibit it again.  that's a hard test to run.19:32
*** reed has quit IRC19:33
sdaguemordred: also, you problem only showed up in neutron job, right?19:33
*** CaptTofu has joined #openstack-infra19:34
openstackgerritA change was merged to openstack-infra/reviewstats: Move openapproved into the reviewstats namespace.
openstackgerritA change was merged to openstack-infra/reviewstats: Move reviews_for_bugs into reviewstats.
openstackgerritA change was merged to openstack-infra/reviewstats: Move openreviews into the reviewstats namespace.
openstackgerritA change was merged to openstack-infra/reviewstats: Move into the reviewstats namespace.
*** sarob_ has quit IRC19:35
*** CaptTofu has quit IRC19:36
openstackgerritA change was merged to openstack-infra/config: Pass $mysql_password through to gerrit class
*** CaptTofu has joined #openstack-infra19:36
fungijeblair: yeah, it was the merging of which triggered that reload... changed a job match from a regex to a literal string, but that's about the only thing there which strikes me as even remotely extraordinary19:36
mordredsdague: I did not - I got distracted/pulled away19:36
mordredsdague: it was the postgres job iirc19:37
sdaguemordred: ok, yeh, that's the 2nd most common job it runs in19:38
jog0sdague: so one of the outcomes of the gate issues was we want qa folks to be able to set any gate bug as critical19:40
jog0accross all projects19:40
jog0so what team should have that access, or do we just make all bug triage teams open?19:40
sdaguejog0: well we need to change permissions to make that happen19:40
jog0sdague: right, thats why I am bringing this up in infra19:40
sdaguehonestly I think all bug triage teams should be open19:41
jog0sdague: ++19:41
jog0fungi clarkb  etc ^: thoughts19:41
jeblairi think they should be open too.  let's start by just asking the ptls of projects with closed teams.19:42
jog0neutron is closed19:42
jog0and glance too19:42
fungimy thoughts as well19:42
jog0if my memory serves me correctly19:42
sdagueneutron is closed, ttx added me by fiat at some point19:42
jeblairmarkwash: do you mind if we make the glance bug triage team open like most of the other projects?19:42
markwashprobably not?19:43
fungiit's not much of an infra discussion really since this is mainly about setting permissions on lp, but might make a good topic for the project meeting agenda19:43
jeblairmarkwash: excellent choice.  :)19:43
jeblairfungi: lp perms are definitely in-scope for infra19:43
fungijeblair: well, yes the how-to discussion is. as for the permissions discussion, i think we have a stake in it along with the ptls was my point19:44
*** rongze has joined #openstack-infra19:44
jeblairmarkwash: ok, done, let me know if it becomes a problem19:44
fungi(whereby i mean the ptls as spokespeople for the projects they represent, but really the projects, or at least the people involved in bug triage for them)19:45
markwashjeblair: cool19:45
jeblairlet's remember to ask markmcclain when he's online19:46
*** rongze has quit IRC19:49
jeblairpabelanger, mordred: you know we don't have the ability to only run certain jobs post-commit based on file paths, right?  so for when a commit lands, we would have to run deb and rpm build jobs.19:49
jeblairpabelanger, mordred: (the jobs could try to figure out if they actually need to do anything, but they'd at least be triggered)19:50
jeblairfungi: okay to approve new projects, or is manage_projects still broke?19:50
pabelangerjeblair, commits are branch specific. Today debian/sid exists, so I _think_ we could build out triggers to be branch aware, some how invoking the project build jobs19:52
pabelangerRPMs would like in fedora/heisenbug as another example19:53
fungijeblair: i think it was clarkb who expressed concerns last week over approving further project additions. but as far as i know there haven't been any changes to the script that i'm aware of (i expect it still needs refactoring to reduce the number of calls it makes to github)19:53
openstackgerritA change was merged to openstack-infra/git-review: Split git rev-parse --show-toplevel --git-dir on newline
jeblairpabelanger: oh, BRANCHES, not directories.  ok.19:54
pabelangerjeblair, Ya, figured that was better to manage19:54
jeblairpabelanger: we can definitely have specific jobs per-branch.19:55
pabelangerjeblair, I was going to get a thread going on the mailing list on how people see the builds working.  I have pbuilder setup for debian / ubuntu builds, however there might be other ways people want it to function19:59
clarkbfungi: jeblair: my concerns are that I have had to fix repos on the zuul side many times now19:59
clarkbI shpould have spent more of that time fixing manage-projects, but making people happy seemed to be the immediate concern19:59
jeblairclarkb: should we suspend creation of new non-openstack projects until that's fixed?20:00
*** dkranz has quit IRC20:00
fungiUtahDave: were you able to find any good examples of setting up chains of dependent actions using reactors?20:01
openstackgerritClark Boylan proposed a change to openstack-infra/config: Create two classes of d-g jobs.
jog0jeblair: not asking for a review, but want to make sure you are aware of as I expect you to -1 it20:01
clarkbfungi: jeblair ^ first patchset of the job refactor. I am sure I missed something so please review :)20:01
jeblairjog0: why?20:01
clarkbjeblair: that was the direction I wanted to head. As it will reduce churn20:01
openstackgerritA change was merged to openstack-infra/elastic-recheck: add fingerprint for bug #1250168
uvirtbotLaunchpad bug 1250168 in neutron "gate-tempest-devstack-vm-neutron-large-ops is failing" [Critical,Confirmed]
jeblairclarkb: is there a bug for this?20:02
jog0jeblair: 'Always run check-tempest-devstack-vm-full twice20:02
jog0if we don't have enough cloud20:02
jeblairjog0: perhaps mordred should -1 it then.  :)20:03
*** jroovers has quit IRC20:03
jog0mordred: ^ do we have enough cloud20:03
clarkbjeblair: I think so, let me double check20:03
jeblairmordred, jog0: istr mordred saying "not right now, but after the move to the new region" or something like that.20:03
mordredI don't have a problem with running check-tempest-devstack-vm-full twice now20:04
jog0mordred: woot20:04
jog0just wanted to make sure this doesn't break the cloudz20:04
clarkbjeblair: though it needs updatnig20:04
uvirtbotLaunchpad bug 1242569 in openstack-ci "No patchset can be merged into any repository" [Undecided,Confirmed]20:04
mordredespecially if it helps keep races out - since races cause resets - which are substantially more activity than an extra job20:04
jog0sdague: thoughts?20:04
jeblairclarkb: would you mind updating it?  if we're going to block new projects, i'd like to have an up to date bug that's set to critical, and i think you're best equipped to write about the problem.20:05
jog0mordred: thanks20:05
clarkbjeblair: yup updating now20:06
sdaguemordred: so I'm not sure it's going to help as much as you think20:06
jeblairfungi: what work is being done on it?  is salt still the preferred resolution?20:06
*** yolanda has quit IRC20:06
*** DinaBelova has quit IRC20:07
sdagueif these are really low percentage bugs (like 1% occurance rate)20:07
*** rnirmal has joined #openstack-infra20:07
sdagueyou'd need to run the tests 69 times to let them get through less than 50% of the time20:08
*** lcestari has quit IRC20:08
*** DennyZhang has joined #openstack-infra20:08
jeblairmordred: ^ we're heading toward not accepting new non-openstack projects until the multi-system manage_projects races are fixed.20:08
fungijeblair: it was last it got discussed. i was hoping to get some examples of setting cross-minion dependencies on tasks (which it had been suggested was tricky and not well documented). but the more i think about it, i worry that still leaves us some races unless we get very, very granular about which new repositories are ready for zuul to clone20:10
fungiright now it's configuration wide "there was a change" but we could still easily tell zuul "update, there was a change" and it ends up doing so in the middle of a subsequent manage-projects run for an even newer addition20:10
jeblairfungi: well zuul won't clone a repo until its in its config, so that at least can be solved by requiring two changes20:11
fungier, i should say "right now the plan is such that"20:11
jeblairfungi: but there's still a gerrit/cgit race, rightL20:11
*** alexpilotti has quit IRC20:11
fungijeblair: agreed20:11
fungithe cgit race is ultimately due to it keying off the same common list of projects which manage-projects looks at20:12
clarkbjeblair: bug updated20:13
fungijeblair: but maybe a belt-and-braces check/retry loop in the cgit cloning operation could mitigate that20:13
clarkbthe cgit race is not a major problem in practice20:13
clarkbbecause occurence of it is low enough that we don't see it for every new project20:13
clarkbwe are seeing this manage-projects breaking zuul thing for every new project20:13
fungii hadn't seen the zuul-related race crop up in ages, but i guess that's because clarkb was catching all of them20:14
openstackgerritA change was merged to openstack-infra/elastic-recheck: Tell people to do a recheck
fungii was seeing the cgit race on many (if not most) project additions20:14
openstackgerritA change was merged to openstack-infra/elastic-recheck: Swap in better query for 1239856
clarkbin zuul:/root/broken-repos I have collected the repos that result from this bug20:15
clarkbif you are curious to see what it looks like20:15
fungiit is just a missing head symref?20:15
sdaguejog0: you totally blew the queries.yaml format :)20:15
*** dkranz has joined #openstack-infra20:15
clarkbfungi: it is missing that, but git status shows all files as deleted20:16
clarkbfungi: I think that is because it clones the empty repo so the working dir is empty20:16
jog0sdague: yeah I haven't gotten around to updating it, good news was unit tests cought it !20:16
clarkbthen zuul does a HEAD=origin/master and all of the files are still deleted20:16
clarkband zuul can't merge on top of that dirty state20:16
*** SergeyLukjanov has quit IRC20:17
jeblairi think we need to fix or work around both problems20:17
fungiclarkb: ahh, right. slightly uglier than what happens on the git farm since those are just bare repos instead20:18
openstackgerritSean Dague proposed a change to openstack-infra/elastic-recheck: add query for bug #1253896 ssh failure bug
uvirtbotLaunchpad bug 1253896 in tempest "test_minimum_basic_scenario fails with SSHException: Error reading SSH protocol banner" [Critical,Confirmed]
*** ryanpetrello has quit IRC20:19
UtahDavefungi: Sorry, I was at lunch. Just got back.20:19
fungiat this point, any expedient fixes/workarounds are unlikely to come from global dependency-based task orchestration as that's still very much poc20:19
fungiUtahDave: no worries20:19
UtahDavefungi: I need to sanitize an example for you. I'll get that together soon.  today20:20
fungiUtahDave: too awesome--thanks!20:20
openstackgerritClark Boylan proposed a change to openstack-infra/config: Create two classes of d-g jobs.
clarkblets see if jenkins is happier with that patchset20:24
*** yassine has quit IRC20:24
*** dhellmann-afk is now known as dhellmann20:25
dhellmannhi, folks20:25
dhellmannI have a stackforge & question. I'm working on importing stevedore, and trying to figure out the rtfd-id value for it, but I don't see that in the UI anywhere. Is there an easy way to figure out that number?20:26
*** hashar has joined #openstack-infra20:26
annegentledhellmann: HMMMmmm20:27
dhellmannI own the project, and can look at the admin UI, but I just don't see the # anywhere20:27
annegentledhellmann: funny I was just going to ask a question about rtfd too20:27
*** dcramer_ has quit IRC20:28
dhellmannannegentle: ask yours, maybe I can at least help you! :-)20:28
openstackgerritA change was merged to openstack-infra/elastic-recheck: Add query for bug 1177134
uvirtbotLaunchpad bug 1177134 in devstack "[ERROR] /opt/stack/new/devstack/exercises/ Image ami-00000001 not available within 15 seconds" [Medium,Triaged]
*** markmcclain has joined #openstack-infra20:29
clarkbdhellmann: our stackforge docs say you can get the info from your rtfd account20:29
clarkbI haven't ever actually used rtfd, this particular thing was a contribution by another stackforge user20:30
dhellmannclarkb: yeah, I was hoping for something more specific20:30
dhellmannI'm looking at the pages under my account and the admin ui for the project, and everything seems to use the project name20:30
dhellmannclarkb: ok, thanks20:30
fungii may have added the "get that info from your rtfd account" bit because previously the docs didn't say anything about that value at all20:31
*** cody-somerville has quit IRC20:31
fungibut like clarkb, i've not used rtfd and have no account there to confirm20:31
clarkbso project IDs are a thing, now how do we get them20:33
dhellmann looks promising20:33
dhellmannoh, same page :-)20:33
dhellmannit appears I can use the API with the slug, and the return value includes the id20:33
clarkbdhellmann: we might be able to change the jenkins job to use a slug instead of an id20:34
clarkbdhellmann: the job hits an api endpoint iirf20:34
dhellmannclarkb: that's a good idea20:34
dhellmannok, I have the id value20:34
dhellmannclarkb: what would we do with existing projects?20:35
clarkbdhellmann: we would need to convert ids to slugs20:35
clarkband make sure the jenkins job works without the id numbers20:35
*** cody-somerville has joined #openstack-infra20:36
*** cody-somerville has quit IRC20:36
*** cody-somerville has joined #openstack-infra20:36
dhellmannclarkb: that job template is a shell call to curl20:36
dhellmannhaha, using the slug just works20:37
annegentledhellmann: clarkb: so I've got an interesting set of questions about incubation and documentation, have a mo?20:37
dhellmannannegentle: sure20:37
clarkbannegentle: sure20:37
*** herndon has joined #openstack-infra20:38
clarkbdhellmann: cool, in that case I think we should convert the other way and change the variable to be rtfd-project or whatever20:38
annegentledhellmann: clarkb: so one of the graduation criteria we're discussing is "show you're supporting users through docs and"20:38
*** yassine has joined #openstack-infra20:38
annegentleso in the incubation and pre-incubation, would projects point to
annegentleand would we then redirect to eventually?20:38
dhellmannannegentle: +1, that seems like a low effort solution to me20:38
dhellmannsupport is there in the tooling already20:39
annegentleif a project wants to do docs "the OpenStack way" would they be better off building docbook from the start?20:39
annegentleif they want to integrate in our install guide, config guide, admin guide, for example, those are pretty specific20:39
dhellmannfor ceilometer, we updated the pre-release version of our docs on rtfd to say "these are old, go look over here" and then left them online with a version number20:39
dhellmannannegentle: ah, good point20:39
annegentleso I hem and haw about actually incubating nicely -- if I were a writer looking to help out a project, how would I view our incubation process for docs?20:39
openstackgerritSean Dague proposed a change to openstack-infra/elastic-recheck: add query for bug #1253896 ssh failure bug
uvirtbotLaunchpad bug 1253896 in tempest "test_minimum_basic_scenario fails with SSHException: Error reading SSH protocol banner" [Critical,Confirmed]
annegentlealso our info arch is so specific, I worry that docs don't serve projects well (how can we improve?)20:40
annegentlepartially I think, well, the internet, that's just how it works, links get redirected20:40
annegentlepartially I think, do we do OpenStack operators a disservice though20:41
dhellmannit seems like having some docs is usually better than having none, and so allowing makes sense -- but that's a good question about transitioning20:41
annegentledhellmann: yeah it's the not-so-smooth transition I worry about20:42
clarkbwe could publish to both places20:42
clarkbwe have discussed it in the past20:42
dhellmannwith ceilometer, I guess that's in process to some degree, but I'm not sure where ceilometer even fits into the existing manuals (I missed any conversations about that at the summit)20:42
dhellmannclarkb: the issue is the use of different input formats, though20:42
annegentleclarkb: can you publish non-sphinx jobs to rtfd?20:42
clarkboh I am ignoring the docbook docs because tehy are special20:42
annegentledhellmann: so we have a chapter to install ceilometer in the install doc20:42
clarkbI am thinking of the in tree docs20:42
dhellmannannegentle: nice!20:43
annegentleclarkb: right, let's say I have a project that wants to put docbook and wadl in the tree20:43
annegentleclarkb: and, we also have decided that API doc specs will go in the tree20:43
annegentledhellmann: so I'm trying to better formulate "the OpenStack doc way"20:43
clarkbI would probably argue against putting docbook stuff in the python trees20:43
clarkbbecause python has a rich set of documentation tools none of which are built around docbook (it is a technical debt dead end)20:44
dhellmannannegentle: how much of the tool chain for building html from docbook input files depends on the docs repo layout, and how much can it be reused?20:44
annegentleclarkb: please don't :) I've spent a lot of time getting PTLs to agree20:44
dhellmannannegentle: if I had a second repo with my stuff in it, could I build html?20:45
*** denis_makogon has joined #openstack-infra20:45
annegentledhellmann: the pom.xml contains the build instructions and they'd be reusable20:45
dhellmannthe thing that gives us is a tool chain for producing the docs in a consumable format, so we would need to have that20:45
*** rongze has joined #openstack-infra20:46
clarkbannegentle: doing that means managing a new set of dependencies...20:46
annegentledhellmann: it's really also about the domain name, do I need to provide support for say,
clarkbwhich won't tie into setup.py20:46
annegentledhellmann: which we've had in the past20:46
clarkbI dunno, it seems very messy20:46
dhellmannannegentle: that's the next question20:46
annegentleclarkb: ohhh interesting20:46
*** salv-orlando has joined #openstack-infra20:46
dhellmannclarkb: why does the doc build need to tie into
*** DennyZhang has quit IRC20:46
dhellmannat least, why for end-user docs?20:46
annegentleclarkb: oh but we're not going to build these api specs20:47
clarkbdhellmann: because that is how we build docs20:47
annegentleclarkb: they're just there for reference, and for example the Keystone v3 API doesn't build, it just gets "published" with github's renderer20:47
clarkbdhellmann: there has long been an expectation that if you have the source you can run tests and build docs20:47
clarkband do so with simple tox commands that install all of your dependencies20:47
sdagueso if cloud resources are getting strained again, have we thought about bringing back fast fail for requirements and pep8, and kicking those jobs out right away?20:47
annegentleclarkb: so actually, for api specs, no building20:47
annegentleclarkb: but it still does make me think twice about end user docs in repos20:47
annegentleproject repos20:47
sdaguenoticing, for instance - fails on license issue20:48
sdaguebut it's going to consume a bunch of devstack nodes for a while before that feedback comes back to the author20:48
clarkbdhellmann: the one exception to that is libvirt, which nova works around with a fake libvirt.20:49
annegentledhellmann: and another doc item you might know, there's still no support that I know of for inclusion in sphinx/rst -- such as, glue together this deliverable with pieces from multiple repos?20:49
annegentledhellmann: or wouldn't have to be across repos, but across folders in a repo?20:49
*** rongze has quit IRC20:50
annegentledhellmann: cool, I knew you'd know :)20:50
dhellmannannegentle: you can also use wildcards in a toctree directive20:51
annegentledhellmann: that would work nicely for ..include:: json examples20:52
annegentledhellmann: but wonder if it'd render xml request/responses ok20:52
clarkbfungi: jeblair: mordred: sdague: and others is ready for review. The diff is large but the change isn't too scary20:53
dhellmannor "include" with the "literal" flag set20:53
clarkbdhellmann: you may be interested in that too. I see WSME/pecan overloading the branch-designator variable at some point to easily have differently named tempest jobs20:53
clarkbdhellmann: it isn't 100% of the way there for WSME/pecan but it is very close20:54
dhellmannclarkb: I will add it to my list of reviews, thanks20:54
*** dprince has quit IRC20:54
clarkbactually it may just work today20:55
*** ruhe has joined #openstack-infra20:55
clarkbafter that change merges20:55
dhellmannannegentle: I don't see an easy answer to the doc location question. We either require new projects to learn the docbook tools, or we have to come up with a migration plan for moving from sphinx to docbook after or during incubation.20:56
dhellmannannegentle: I would lean toward coming up with a migration plan, even if that is "you'll have to learn to use docbook to graduate"20:56
dhellmannannegentle: because, IMO, that's a lower barrier to getting started on a project -- but maybe we can make the docbook tooling easy enough, too?20:57
fungisdague: i'm getting pretty sure that the check jobs are only contributing a small amount to resource consumption when we get into this state. the bigger issue is that the gate is deep enough that a reset near the head of the shared queue is enough to waste as many devstack nodes as our aggregate quota, so the entire pool is depleted and waiting for more nodes, and it cycles like that for every reset20:57
fungicompounding the delay20:57
*** marun has quit IRC20:57
*** yamahata_ has quit IRC20:58
sdaguefungi: fair20:58
*** yamahata_ has joined #openstack-infra20:58
sdaguethe gate was empty when I sent 25 patches into it this morning :)20:58
annegentledhellmann: honestly, people can write on a napkin :)20:58
clarkbannegentle: dhellmann: we can use markdown with docbook, can we convert rst as well?20:58
clarkbannegentle: or was the markdown experiment a failure?20:58
fungisdague: so you're saying we can blame you this time? ;)20:59
annegentledhellmann: clarkb: markdown and rst are not amenable to "chunking" for uniquely identifying "what is a page" so that comments can be tracked. Only docbook and asciidoc give us that (to my understanding)20:59
clarkbannegentle: really? many static blog generators use both just fine20:59
sdaguefungi: mine have long since cleared20:59
*** nati_ueno has joined #openstack-infra21:00
clarkbI think they use the model of file per thing with comments which may not scale well21:00
sdaguehonestly, our throughput isn't too bad today21:00
*** jorisroovers has joined #openstack-infra21:00
fungisdague: agreed. the head of the integrated queue was approved a little over 4 hours ago21:01
clarkbannegentle: is an example of that21:01
*** sarob has quit IRC21:01
openstackgerritJoe Gordon proposed a change to openstack-infra/elastic-recheck: Add query for 1254872
*** emagana has joined #openstack-infra21:02
*** sarob has joined #openstack-infra21:02
annegentleclarkb: too weird that pelican is like publican21:02
annegentleclarkb: and yeah, it requires page-per-file. we no like21:03
*** dhellman_ has joined #openstack-infra21:03
*** dhellmann has quit IRC21:03
openstackgerritA change was merged to openstack-infra/elastic-recheck: add query for bug #1253896 ssh failure bug
uvirtbotLaunchpad bug 1253896 in tempest "test_minimum_basic_scenario fails with SSHException: Error reading SSH protocol banner" [Critical,Confirmed]
annegentleclarkb: redhat builds with publican21:03
* dhellman_ is having connectivity issues21:03
*** jorisroovers has quit IRC21:05
*** jorisroovers has joined #openstack-infra21:05
*** flaper87 is now known as flaper87|afk21:06
*** sarob has quit IRC21:06
*** mihgen has quit IRC21:07
*** dkranz has quit IRC21:08
*** marun has joined #openstack-infra21:09
clarkbannegentle: is that something that could be used to do a migration though?21:09
clarkblive with a few extra files until they can be converted properly to docbook21:09
annegentleclarkb: yeah migrations not difficult if they understand what we want in the end (identifiable chunks). Probably matters to translation too21:10
annegentleclarkb: dhellman: and another thing I was pondering on, you might have ideas about,21:11
annegentleclarkb: dhellman_ from start to finish, an incubation > integrated might take a year?21:11
annegentleclarkb: dhellman_ is that right?21:11
annegentleclarkb: dhellman_ : or is it more like six or seven months21:11
clarkbI think it is usually at least one cycle21:12
annegentleclarkb: yeah that's what I was thinking too21:12
*** wayneeseguin has quit IRC21:13
annegentleclarkb: so if the TC votes next week to incubate a project, they could be part of J release as integrated?21:14
denis_makogongoodday, -infra21:14
denis_makogoni've got few questions21:14
clarkbannegentle: I think in theory yes, in practice I doubt we would see that happen, just due to the way things have gone previously21:14
denis_makogonit would be nice to hear answers from anyone21:14
annegentleclarkb: yeah me too21:15
annegentledenis_makogon: go ahead and ask21:15
annegentledenis_makogon: seems interruptive but it's not really :)21:15
denis_makogoni've came from Trove community21:15
dhellman_annegentle: so far we've only done the transition in one cycle, but I think with some of the new rules we are going to see multiple cycle incubations21:15
dhellman_annegentle: and I think that's perfectly fine and we just need to set the expectations of new projects accordingly21:15
denis_makogoni've need to add some code to trove gate21:15
denis_makogonto unable new tests21:15
annegentledhellman_: yeah this is all related to our discussions21:15
dhellman_annegentle: yeah, I need to read those etherpads21:16
denis_makogonhow can i do that ?21:16
annegentledhellman_: but also trying to manage my stress level with lots of people asking about docs for incubation21:16
clarkbdenis_makogon: which test(s) are you looking to update?21:16
clarkbdenis_makogon: specifically which jenkins jobs that report back21:16
*** wayneeseguin has joined #openstack-infra21:16
*** ruhe has quit IRC21:17
*** wayneeseguin has quit IRC21:17
*** wayneeseguin has joined #openstack-infra21:17
denis_makogonclarkb, i need to update whole build(reddwarf) process21:17
*** dolphm has joined #openstack-infra21:17
clarkbdenis_makogon: right, but in which jenkins jobs?21:18
dhellman_annegentle: do you want integrated projects to be included in the regular manuals, or are you planning to have separate manuals for things that are not in whatever we're calling the core set?21:18
*** rfolco has quit IRC21:18
annegentledhellman_: so I had to draw the line even in our "mission statement" that we only support core with the resources we have21:18
annegentledhellman_: but, we managed to somewhat doc heat and ceilometer in the install guide21:18
dhellman_annegentle: perfectly reasonable21:18
*** dcramer_ has joined #openstack-infra21:18
annegentledhellman_: so I'm looking for ways to get good quality "OpenStack" docs while also paying attention to individual projects21:19
dhellman_annegentle: I think it's perfectly reasonable to say that someone from the project needs to work with you guys and submit patches21:19
annegentledhellman_: incubated ends up wayyy down the priority list, but I'm wondering, if we had a better pipeline, would we get better results21:19
dhellman_just like we say we want tempest tests21:19
annegentledhellman_: that's a good analogous resource21:20
*** CaptTofu has quit IRC21:20
dhellman_annegentle: part of it, for me, is that the setup is "mysterious" so maybe its a (excuse the pun) doc issue? or maybe I just haven't read the right thing?21:20
hub_capclarkb: we are looking to put our oneoff tests and setup into a job that gates only trove21:20
*** CaptTofu has joined #openstack-infra21:20
dhellman_annegentle: the latter is very likely, regardless of the former21:20
annegentledhellman_: possibly, though I also get input that our HowTo page is great21:20
clarkbhub_cap: one off meaning the third party tests?21:20
dhellman_annegentle: ok, I may just not have found that then21:21
hub_capmeaning proboscis tests clarkb21:21
dhellman_annegentle: let's say I need education :-)21:21
annegentledhellman_: it's as if it's mysterious if you're a dev  because of xml allergies :)21:21
hub_capoh yea21:21
*** wayneeseguin has quit IRC21:21
dhellman_annegentle: I'm not allergic to xml21:21
clarkbhub_cap: denis_makogon: what sort of environment does proboscis expect to run the tests in?21:21
annegentledhellman_: it's long though.21:21
clarkbthis isn't normal unittesting is it?21:21
hub_capclarkb: we run them on a vanilla ubuntu vm21:22
annegentledhellman_: for sure we're working on a doc that says clearly where to put API docs21:22
*** wayneeseguin has joined #openstack-infra21:22
*** dkranz has joined #openstack-infra21:22
annegentledhellman_: but maybe the rest of the docs are mysterious because they're relatively new21:22
hub_capclarkb: not @all. its full blown integration stuff21:22
dhellman_annegentle: maybe the thing to do is add some template files to the cookiecutter repository?21:22
annegentledhellman_: so yeah definitely need to make sure people know what titles we have21:22
annegentledhellman_: yes, we've had those21:22
dhellman_annegentle: knowing where I would need to add something is another question21:22
annegentledhellman_: but I don't want more books :)21:22
dhellman_annegentle: cool, I'm just behind21:22
annegentledhellman_: I only want more pages21:23
annegentledhellman_: does that make sense?21:23
clarkbhub_cap: denis_makogon: I see. For that you will want to overload the devstack-gate process to run your stuff instead of tempest. Swift and pbr have both done that effectively. I will get you links to their job configs21:23
annegentledhellman_: people need to add to existing titles21:23
dhellman_annegentle: well, yeah, but if I'm going to test what I write, I need enough of a project to build it21:23
annegentledhellman_: true21:23
hub_capclarkb: <321:23
*** rfolco has joined #openstack-infra21:23
denis_makogonclarkb, <#21:23
annegentledhellman_: but I'd prefer you write within the titles we have, does that cause difficulty?21:23
*** wayneeseguin has quit IRC21:23
clarkbhub_cap: denis_makogon: is what swift does21:23
*** sarob has joined #openstack-infra21:23
mordredhub_cap: but don't think it's going to make me stop smacking you in the face about it21:23
clarkbhub_cap: denis_makogon: swift is using devstack to install a full stack then running their functional tests against that21:24
*** CaptTofu has quit IRC21:24
portanteclarkb: gate queue seems kinda long, need any help on the swift side?21:24
annegentledhellman_: hmmm possibly I'm asking too much of people -- to know all the titles we have21:24
*** CaptTofu has joined #openstack-infra21:24
dhellman_annegentle: what if I create mini versions of those titles, with a single chapter that can then be incorporated into the official guide when my project graduates?21:24
annegentledhellman_: bingo, that's what I'm thinking of for incubated21:24
hub_capmordred: worth a try, tho :)21:24
clarkbportante: I don't think so, sdague mentioned earlier that throughput is good, I think we may just be running into more backlog from last week21:24
clarkbportante: thank you thouhg21:24
dhellman_annegentle: it might be easier to say one guide with a chapter for each title you have, but as you say, the project team needs to know what those titles are21:25
*** wayneeseguin has joined #openstack-infra21:25
mordredhub_cap: is pbr doing a similar thing to swift - note in the pbr one you could not include in the gate_hook if you did not want to run devstack21:25
mordredhub_cap: how do your tests work? like, where do they get nova from?21:25
clarkbif you don't run devstack you will be ersponsible for a bit of setup, but there is nothing stopping you from doing that21:25
hub_capmordred: they use a script to download / spin up devstack mordred21:26
hub_capand it builds an image and creates some users and downloads the extra things needed like proboscis etc..21:26
mordredok. so just having a devstack there already could potentially be helpful then21:26
*** julim has quit IRC21:26
hub_capmordred: most def21:26
mordredgood. btw, you need to delete this:
* mordred punches hub_cap in the face repeatedly21:27
hub_capyes mordred once designate support is added we can deprecate and remove it in N+N^2 cycles21:27
annegentledhellman_: clarkb thanks much you guys are the best21:27
mordredhub_cap: you can remove it in 0 cycles. BOOM21:27
*** MarkAtwood has joined #openstack-infra21:28
mordredhub_cap: once designate support is added by whom?21:28
dhellman_annegentle: I'm glad I could help!21:28
hub_capmooooooom montys being meeannnnnnnn21:28
hub_capmordred: hp has a review for it21:28
denis_makogonmordred, and it's landed into codebase21:28
*** hashar is now known as hasharCall21:28
mordredhub_cap: srrsly. it's really weird to see an openstack project depend on code to be able to talk to a non-openstack single-cloud service21:29
esmute@hub_cap @mordred: ALready merged.
mordredthat makes me happy21:29
*** wayneeseguin has quit IRC21:29
mordrednow I guess we should at some point get designate incubated or something :)21:29
*** dcramer_ has quit IRC21:30
hub_capso mordred in HH we can remove it (/me assumes we are doing vending machine alphabet for the openstack releases)21:30
*** wayneeseguin has joined #openstack-infra21:30
denis_makogonmordred, sounds good, because since trove support designate we need to add functional tests to see if it works fine21:30
*** melwitt has quit IRC21:31
*** melwitt1 has joined #openstack-infra21:31
openstackgerritJoe Gordon proposed a change to openstack-infra/elastic-recheck: Add query for 1254872
clarkbhub_cap: I want us to start using unicode characters21:34
jeblairclarkb: we are using unicode characters21:35
*** weshay has quit IRC21:35
clarkbjeblair: fair, but more of the thousands (millions?) than the 26 available to us with the alphabet21:35
jeblairspeaking of which, i note that rackspace misspelled f19's name.21:35
clarkbthe 27th release can be snowman21:36
hub_capclarkb: ya, i bet it'd be hard to find a HH town close to the summit21:36
dhellman_hey, guys, if I want a stackforge project to be able to publish releases to an existing project on pypi, which user do I need to add permissions for?21:37
*** dhellmann has joined #openstack-infra21:37
jeblairdhellman_: openstackci21:37
dhellman_jeblair: thanks21:37
openstackgerritDoug Hellmann proposed a change to openstack-infra/config: Move stevedore to stackforge
*** dhellman_ has quit IRC21:40
*** fifieldt has joined #openstack-infra21:42
jeblairclarkb: thanks for writing up bug 124256921:43
uvirtbotLaunchpad bug 1242569 in openstack-ci "manage-projects error on new project creation" [Critical,Triaged]
openstackgerritA change was merged to openstack-infra/elastic-recheck: Add query for 1254872
jeblairnow we need someone assigned to it;  it looks like clark has some non-salt possibilities for fixing it.  any volunteers?21:44
*** dhellmann is now known as dhellmann-afk21:45
clarkbI am no longer up to date on manage-project's control flow after recent changes. I can look at it, but I think mordred is most familiar with it21:46
jeblairmordred, clarkb, fungi: i'm going to start -2ing changes that add projects and reference that bug.21:46
clarkbmordred: maybe you can read the bug comments for 1242569 and at least confirm we are on the correct track then someone else can do the work21:46
mordred21:45:11 MarkAtwood | he gets 5 people, so its being an agonizing game of lifeboat-captain for him21:46
mordred21:45:35    mordred | he doesn't realize that the rest of HPC isn't an agonizing place right?21:46
*** rongze has joined #openstack-infra21:46
mordredI hate paste errors21:46
*** ruhe has joined #openstack-infra21:46
*** jhesketh has joined #openstack-infra21:47
mordredjeblair: looking at bug21:47
openstackgerritClay Gerrard proposed a change to openstack-dev/hacking: Add noqa support for H201 (bare except)
*** ruhe has quit IRC21:48
*** yassine has quit IRC21:49
*** mrodden has quit IRC21:49
mordredclarkb: thank you. that's a very helpful writeup21:51
mordredclarkb: I'll work on a fix - but I probably won't get to it til later tonight or tomorrow21:52
*** mrodden has joined #openstack-infra21:53
*** rongze has quit IRC21:54
lifelessmordred: fun quote there :P21:54
mordredlifeless: shush21:54
*** dolphm has quit IRC21:57
*** dolphm has joined #openstack-infra21:57
*** rfolco has quit IRC21:58
clarkbjeblair: I have been copy pasting a note into the new d-g job chagnes linking back to my change with a small description of what is going on. Now that we have that it should be much easier toget everyone on the same page21:59
*** mfer has quit IRC22:00
*** ^d is now known as ^demon|meeting22:01
*** Ryan_Lane has quit IRC22:02
openstackgerritA change was merged to openstack-infra/config: Remove undefined jobs and update other references
openstackgerritA change was merged to openstack-infra/config: Remove jeepyb from the requirements gate
*** nati_ueno has quit IRC22:06
lifelessclarkb: needs second +222:06
*** UtahDave has quit IRC22:07
*** nati_ueno has joined #openstack-infra22:07
*** Ryan_Lane has joined #openstack-infra22:07
*** ArxCruz has joined #openstack-infra22:08
*** reed has joined #openstack-infra22:11
*** ruhe has joined #openstack-infra22:14
*** pcm_ has quit IRC22:15
*** Ryan_Lane has quit IRC22:15
*** senk has quit IRC22:15
jeblairclarkb: is affected by your refactor?22:15
jeblairclarkb: and should we require non-integrated-gate devstack jobs be in a different file?22:16
jeblairmordred: i have assigned bug 1242569 to you based on "I'll work on a fix"22:17
uvirtbotLaunchpad bug 1242569 in openstack-ci "manage-projects error on new project creation" [Critical,Triaged]
clarkbyes 57098 is affected, I will comment on that. I like putting things that use d-g hooks in other files, like requirements and swift functional tests22:18
jeblairclarkb: okay, i will comment with the request to move to another file22:18
openstackgerritA change was merged to openstack-infra/git-review: Remove empty lines from git log output
jeblairclarkb: also touches dg files22:20
*** fbo_away is now known as fbo22:21
*** Ryan_Lane has joined #openstack-infra22:23
openstackgerritKhyati Sheth proposed a change to openstack/requirements: Add dependency django-clippy
*** dcramer_ has joined #openstack-infra22:26
*** sandywalsh has quit IRC22:26
openstackgerritA change was merged to openstack-infra/jeepyb: Remove run-mirror from jeepyb
openstackgerritA change was merged to openstack-infra/jeepyb: Only push tags to gerrit
openstackgerritArx Cruz proposed a change to openstack-infra/config: Adding nodepool compatibility with Fedora.
fungistepping out to grab a bite, but back shortly to continue the excitement22:29
*** Ryan_Lane has quit IRC22:30
lifelessjog0: hey22:31
*** dangers is now known as danger_fo_away22:31
*** ruhe has quit IRC22:31
lifelessjog0: so I'm looking at
lifelessjog0: from
lifelessjog0: whats the way - or the docs on the way - to match this against a recheck ?22:31
clarkbjeblair: commented on that one too22:32
*** ryanpetrello has joined #openstack-infra22:32
lifelessclarkb: if you're handing out reviews -
uvirtbotLaunchpad bug 1177134 in devstack "[ERROR] /opt/stack/new/devstack/exercises/ Image ami-00000001 not available within 15 seconds" [Medium,Triaged]22:35
lifelessjog0: ok, so you're a *great* black box tool.22:35
lifelessjog0: now, how do I do that without pinging you?22:35
jog0that didn't trigger elastic-recheck because e-r doesn't know about grenade yet22:35
jog0lifeless: so I don't have a great answer for you yet22:35
lifelessjog0: ah, so you just googled for the bug ?22:36
jog0either teach e-r about grenade22:36
jog0lifeless: you can search openstack bugs22:36
jog0but I saw that recently here:22:37
*** loq_mac has joined #openstack-infra22:37
clarkbteaching e-r about grenade isn't super hard, keeping the number of http GETs that get 404s to a minimum is hard. I could special case grenade I guess22:37
clarkbjeblair: would you be ok with special casing grenade in the confings?22:37
jog0sdague: you had a WIP on this22:37
pabelangerSo, I'm tracking down a issue in #puppet-openstack about a mismatch of the git repo and what is publish on with that in mind, what are you're thoughts about creating puppet module tarballs (via jenkins) and publishing them to tarballs.o.o? Another process could be put in place to upload them into forge.puppetlabs.com22:37
* jog0 goes off to find a late lunch22:38
jeblairclarkb: how so?22:38
*** dkranz has quit IRC22:38
*** melwitt1 has quit IRC22:38
*** melwitt has joined #openstack-infra22:38
clarkbjeblair: the grenade logs go under new/logs old/logs or some such. To get all of those indexes I would triple the number of files configured in the gearman client22:38
clarkbor I would have to teach the gearman client to parse apache index listings to determine which files are available22:39
jeblairpabelanger: i think we'd love that.  that's how we publish to pypi and publish jenkins modules to maven repos.22:39
pabelangerRight now it looks to be a manual process22:39
clarkbjeblair: the simple fix is to make it only look in those dirs if the job matches .*grenade.*22:39
pabelangerand I have something inhouse today that automates based on the existing -infra logic22:39
jeblairpabelanger: the main tech requirement there is that the actual upload needs to run without executing any user-submitted code (eg, with wget, something similar, or a very simple script that isn't kept in the same repo)22:39
*** mriedem has quit IRC22:39
*** markmcclain has quit IRC22:40
jeblairpabelanger: (because it's run on a slave with sensitive creds)22:40
pabelangerjeblair, understood22:40
pabelangerthat will likely require some brainstorming from peeps here, but I think I can get the building tarballs hammered out pretty easily \22:41
*** rfolco has joined #openstack-infra22:41
jeblairclarkb: when people have been saying "logstash does't support grenade" they mean "logstash doesn't support any logs from any grenade jobs", not "logstash doesn't support grenade.log" (or whatever) ?!22:42
clarkbjeblair: it supports console.html but all of the other logs are hidden from it due to paths22:42
clarkbjeblair: we can parse the files just fine, we just don't know where to find them. I have deferred this problem bceause on the surface it is simple but you end up with lots of GETs or workarounds22:43
jeblairclarkb: my entire world view has shifted.  i had no idea that was the problem based on the problem statement.22:43
*** sandywalsh has joined #openstack-infra22:43
clarkbjeblair: :/ hopefully it is more clear now22:43
jeblairclarkb: what do you want logstash to do with the two copies of, say, the nova api log?22:44
*** mrodden has quit IRC22:45
clarkbjeblair: the full location path of the log is indexed too, so having two copies in different places is fine22:45
clarkbjeblair: you can still differentiate them. They will need to be processed as individual files like the other files22:45
jeblairclarkb: roger.  okay, so what's wrong with lots of gets?22:45
clarkbjeblair: we already do ~30 for each test completion. This would bump it to ~9022:46
jeblairclarkb: meh.  :)22:46
jeblairclarkb: apache is fast with 404s.22:46
clarkbin that case, anyone can propose that change to I probably won't get that done today as I am just about to head out fo a late lunch then do some portlandy things22:47
*** rfolco has quit IRC22:47
clarkbjust need to add the new loactions, you can copy pasta the tags and other args for each file22:48
jeblairclarkb: if the config will get yucky, that might be a reason to handle it in code22:48
clarkbthe config will get a bit unweildy22:49
lifelessclarkb: / jeblair: / mordred: I would appreciate your gear insight -
clarkbI think best thing may be to have the config note file names we care about and various attributes, very similar to what we have now. Then do a GET for the index listing and only grab the files that are present22:49
clarkbI will need to fiddle with that though22:49
*** ryanpetrello has quit IRC22:49
*** rfolco has joined #openstack-infra22:50
jeblairclarkb: ok.  i'm more concerned with the config than the GETs, fwiw.22:50
*** rongze has joined #openstack-infra22:50
*** xeyed4good has joined #openstack-infra22:51
*** xeyed4good has left #openstack-infra22:51
jeblairlifeless: ack22:51
*** mrodden has joined #openstack-infra22:53
*** hasharCall has quit IRC22:54
mordredjeblair: ack on the bug - I'm going to dinner, will work on it when I get back22:54
dimsweird, seeing this in a job that just failed22:54
dims2013-11-25 22:51:37.187 | 2013-11-25 22:49:56 error: Installed distribution iso8601 0.1.4 conflicts with requirement iso8601>=0.1.822:54
mordredjeblair: I think clarkb's suggestion of sequencing should be easy enough to encode22:54
*** rongze has quit IRC22:55
* clarkb is going to AFK as well. lunch time22:56
*** heyongli has joined #openstack-infra22:57
*** denis_makogon has quit IRC22:58
*** julim has joined #openstack-infra22:58
*** nati_ueno has quit IRC23:02
*** vipul is now known as vipul-away23:03
*** ^demon has joined #openstack-infra23:03
*** eharney has quit IRC23:06
*** ^demon|meeting has quit IRC23:07
*** julim has quit IRC23:09
harlowjaanyone got some time for a question about the ML and DKIM23:13
jeblairharlowja: go for it23:13
harlowjaso i've been told that y! is turning on DMARC (guess it didn't have it before) but of course it seems like this will affect the ML message that i sent to the ML23:14
harlowjajust wondering if/how others are dealing with this23:14
harlowjasupposedly once this turns on (soon) any mails i sent to openstack-* ML will start to get rejected, unless the ML changes how it works (resigns them maybe?)23:15
*** nati_ueno has joined #openstack-infra23:16
harlowjawas wondering if there is anything in the ML software that can help here (?)23:16
*** ^demon has quit IRC23:16
*** sarob has quit IRC23:18
*** julim has joined #openstack-infra23:19
*** sarob has joined #openstack-infra23:19
jeblairharlowja: thinking23:20
*** bpokorny has quit IRC23:20
harlowjait appears that if DKIM is added, and if the ML forwards it on, then it will bounce on the reciever side23:20
* harlowja not so sure on the technical details really 23:21
openstackgerritRussell Bryant proposed a change to openstack-infra/config: Add gate-solum-devstack job
*** vipul-away is now known as vipul23:22
*** sarob has quit IRC23:23
*** heyongli has quit IRC23:23
wenlock_harlowja, i wonder if adding the ML to your contacts would avoid bounces or sending it to junk23:24
harlowjait seems like from my discussion that unless the ML actually removes the DKIM verification that it will bounce before it even gets to your inbox23:24
harlowjaor if the ML rewrites the DKIM (but not using the senders address)23:24
harlowjathen that might work23:24
harlowjathe later makes it hard to know who sent the email (since the current ML actually proxies as the original user)23:25
harlowjathe actual recommendation i got was to send to the ML from a new different address, lol23:26
*** thomasem has quit IRC23:26
harlowjabut was wondering if the ML software can do better23:26
jeblairharlowja: based on  ...  i'm a bit concerned23:26
jeblairharlowja: it's easy enough for us to configure mailman to stop altering messages (i'd love a good excuse to do that actually)23:27
jeblairharlowja: so that they don't break dkim sigs23:27
jeblairharlowja: but 'identifier alignment' could be tricky23:27
harlowjajeblair ya, seems like more companies are actually turning this on also (so that someone can't forge an email from yahoo-inc)23:27
*** dolphm has quit IRC23:28
*** rfolco has quit IRC23:28
jeblairharlowja: because if dmarc expects the envelope sender (which should be something@lists.o.o) to either match the From address (, or match the yahoo-inc SPF rules (which are likely not going to allow envelope senders
jeblairharlowja: then I don't believe even their first suggestion ('strict forwarder') is possible23:29
harlowjaya, hmmm, i wonder about 'Original Authentication Results'23:30
harlowjafrom internal email23:30
harlowjaDMARC works like this:23:30
harlowja1) We DKIM sign mail.23:30
harlowja2) We tell the world we DKIM sign our mail.23:30
harlowja3) When somebody receives mail that claims to be from us but is not DKIM signed, they reject it.23:30
harlowjaMailing lists that forge From: lines cause 3) to happen when subscribers receive mail. This enforcement happens at the receiving end.23:30
harlowjaMailing lists can change configurations so they're not using the sender's From: line, or receivers can avoid enforcement on mailing lists, but most of them don't.23:30
jeblairi don't know what they mean by "forge From: lines".23:31
harlowja"I know of 0 mailing lists that preserve DKIM. If the mailing list doesn't use the sender address in the From:, it will work correctly."23:31
clarkbclearly we need wave to happen but its not going to happen23:31
jeblairwe don't forge from lines if you send a message, we leave your From address in there.  but perhaps they mean not changing From is forging.23:32
*** thedodd has quit IRC23:32
harlowjaya, i can ask for more details there23:32
jeblairharlowja: I honestly don't know how to parse that sentence; it would be nice if it specifically described what the author was envisioning.23:33
harlowjanp, jeblair asking for more info23:34
clarkbgoogle wave now an apache thing23:34
jeblairharlowja: it sounds like with this level of vagueness, it would be a good idea to be prepared to move to another address though...23:34
*** yamahata_ has quit IRC23:34
*** ryanpetrello has joined #openstack-infra23:34
jeblairharlowja: however, when it is turned on and we can see exactly what has been configured, i'm certainly willing to experiment and look into seeing if there's a workable mailman configuration23:35
zaroclarkb: isn't wave the same as etherpad?23:35
harlowjajeblair sounds good23:35
clarkbzaro: no it is a lot more23:35
harlowja jeblair (found this)23:35
harlowjai think the option 'rewrite From: email header field to contain the list domain in the From header field. e.g.' is what was being talked about in the forging (maybe)23:37
reedclarkb, how can wave happen?23:38
clarkbI don't think it can :/ is it even under active dev?23:38
reedclarkb, oh, then we're on the same page :) I think that train has passed23:39
reedclarkb, seems a better cause23:39
jeblairharlowja: i wonder what configuration uses.  :)23:40
harlowjagood question jeblair23:40
harlowjau'd hope they would get it right23:40
*** ryanpetrello has quit IRC23:40
jeblair(i have subscribed so maybe i'll get an email sometime)23:40
harlowjai'll let u know if i get any more details23:41
*** nicedice has quit IRC23:41
harlowjamaybe its something simple to enable23:41
reedi'm freaking out reading about this dmarc thing23:41
*** nicedice has joined #openstack-infra23:42
reedsounds like an effort by $BIG_US_CORP to decide who can exchange emails with whom23:42
harlowjafrom what i've been told they want to turn it on so that people can't act like they are sending things from yahooo-inc23:43
harlowjathe spoofing part23:43
reedand it's all our fault, for failing to make OpenPGP or similar easy to use23:43
reedharlowja, I understand the purpose ... nonetheless I see how it can be used to exclude from email those that don't have DMARC and I freak out23:44
* reed running around like headless chicken yelling at cospiracy23:44
harlowja'Facebook, LinkedIn, and Twitter made -- tell employees to move their mailing list subscriptions. (Google also moved them, although to another domain.)' ...23:44
*** ^d has joined #openstack-infra23:44
*** vipul is now known as vipul-away23:45
*** rcleere has quit IRC23:45
harlowjajust seems like why move them if the ML software can be configured correctly23:45
reedlet me guess: move to google groups?23:45
harlowjai reacted internally by like saying WTF23:45
harlowjabut apparently this DKIM stuff is actually getting used, which means ML are problematic i guess23:46
jeblairreed: no he's saying bigcorps tell their employees to subscribe to MLs from a different, non-bigcorp address23:46
jeblair(i think)23:46
harlowjajeblair afaik thats a yes23:46
*** jgrimm has quit IRC23:46
reedoh, that makes sense23:46
harlowjawhich seems like ok, sure, but why23:46
harlowjamy guess is more big/smallcorps will start doing this (if they aren't already)23:47
harlowjaso if mailman supports something already...23:47
reedharlowja, I wonder if my email provider supports that, too23:50
harlowjaapparently gmail and y!mail do (i think)23:51
*** rongze has joined #openstack-infra23:51
harlowjaso if everyone starts supporting this, mailman supporting it will become a big issue23:51
*** ruhe has joined #openstack-infra23:52
*** ruhe has quit IRC23:53
harlowjaand apparently if dkim is stripped by mailman then certain people won't recieve it, so its like a crappy situation all around23:54
reedmy provider, 1and1 doesn't seem to support it23:54
*** rongze has quit IRC23:55
jeblairharlowja: i'm pretty sure dkim isn't an issue (or wouldn't be if we set the right options in mm), only the additional bits in DMARC about senders.23:55
harlowjakk, could be jeblair23:56
*** fbo is now known as fbo_away23:57
*** loq_mac has quit IRC23:57
*** nati_ueno has quit IRC23:58
*** oubiwann has quit IRC23:58
*** adarazs has quit IRC23:59
*** nati_ueno has joined #openstack-infra23:59
*** emagana has quit IRC23:59

Generated by 2.14.0 by Marius Gedminas - find it at!