Tuesday, 2014-03-25

Lotus907efiah, ok00:00
*** Hefeweizen has joined #tripleo00:02
openstackgerritlifeless proposed a change to openstack/tripleo-image-elements: Permit overriding the nova compute manager.  https://review.openstack.org/8267700:06
openstackgerritlifeless proposed a change to openstack/tripleo-image-elements: Fixup Ironic support.  https://review.openstack.org/8162700:06
Lotus907efihmmm, I see different behavior if I have DIB_DISTRIBUTION_MIRROR set once the run get to the end user test image than if I do not have it set00:09
Lotus907efiif DIB_DISTRIBUTION_MIRROR then the end user test image seems to be built using external sources00:09
Lotus907efiwait, if DIB_DISTRIBUTION_MIRROR is NOT set then the end user test image seems to be built using external sources00:10
openstackgerritJames Slagle proposed a change to openstack/tripleo-image-elements: Add fedora-rdo-icehouse element  https://review.openstack.org/8065800:10
Lotus907efiif it IS set then the end user test image seems to be built using my internal mirror but things fail due to path differences00:10
*** epim has quit IRC00:10
Lotus907efiI think because if DIB_DISTRIBUTION_MIRROR is set and it gets to building the end user test image, the special sources.list is not used00:11
Lotus907efiand thus the paths to the files are not right00:11
Lotus907efidoes that sound plausible?00:15
lifelessit sounds like your setting for DIB_DISTRIBUTION_MIRROR may be wrong00:15
lifelessthat setting is handled by the ubuntu/fedora etc elements00:15
Lotus907efiso perhaps I should unset that and just make the addition of the apt-sources element to the end user test image build line00:17
openstackgerritlifeless proposed a change to openstack/tripleo-heat-templates: Fix Ironic templates  https://review.openstack.org/8241400:17
lifelessLotus907efi: well what errors do you see?00:17
lifelesswith DIB_DISTRIBUTION_MIRROR and the user image?00:18
Lotus907efierrors about not being able to find the right Package files and etc in the repo00:18
*** openstack has joined #tripleo00:20
-dickson.freenode.net- [freenode-info] channel flooding and no channel staff around to help? Please check with freenode support: http://freenode.net/faq.shtml#gettinghelp00:20
Lotus907efiI think I will try a run with no DIB_DISTRIBUTION_MIRROR set but add the apt-sources element to the end user test image00:20
Lotus907efiand see what happens00:20
lifelessLotus907efi: sure, but you said path errors00:21
lifelessLotus907efi: and I want to see the error you se00:21
Lotus907efithey are in the pastebin I posted earlier00:21
openstackgerritlifeless proposed a change to openstack/tripleo-image-elements: Fixup Ironic support.  https://review.openstack.org/8162700:22
openstackgerritlifeless proposed a change to openstack/tripleo-image-elements: Permit overriding the nova compute manager.  https://review.openstack.org/8267700:22
Lotus907efii.e. the pulls of various pieces from the ubuntu mirror are not found in the right path under the DIB_DISTRIBUTION_MIRROR URL00:22
Lotus907efinot shell path problems00:23
lifelessLotus907efi: the pastebin you pasted showed archive.ubuntu.com being used00:24
lifelessLotus907efi: is that what you mean by a path problem ?00:24
Lotus907efihmm, no00:24
Lotus907efilet me post a different one00:24
*** matsuhashi has joined #tripleo00:26
*** howleyt_ has quit IRC00:27
Lotus907efilifeless: see the end of this log http://paste.openstack.org/show/74205/00:28
lifelessand does exist ?00:33
*** fandi has quit IRC00:34
*** CaptTofu has joined #tripleo00:34
Lotus907efiyes and no, it exists further down in the directory tree00:34
Lotus907efiour sources.list file looks like this:00:35
Lotus907efideb saucy main restricted multiverse universe00:35
Lotus907efideb saucy-updates main restricted universe multiverse00:35
Lotus907efideb saucy-security main restricted universe multiverse00:35
tchaypohaving read conversation above I'm going to skip searching for anything else and go straight to booking the hilton doubbletree via cwt00:36
lifelessLotus907efi: looks like you've a broken mirror00:36
lifelessLotus907efi: I have a local mirror00:36
lifelessexists in it00:37
Lotus907efiit works fine with the seed, undercloud, and overcloud images00:37
lifelessLotus907efi: you are perhaps not mirroring saucy-security00:37
Lotus907efiyes it is there00:37
lifelessLotus907efi: which you're using apt-sources.list00:37
StevenKtchaypo: That was my plan00:37
lifelessLotus907efi: so the DIB_DISTRIBUTION_MIRROR vaue is irrelevant for those00:37
lifelessLotus907efi: value00:37
Lotus907efiso if I get the end user test image to use apt-sources it should work00:37
lifelessLotus907efi: I'm confused, you said 'yes and no'00:38
StevenKAt the moment, I'm trying to avoid 7 hours in LAX.00:38
lifelessLotus907efi: is it 'yes', it exists, or 'no, it does not'00:38
lifelesshttp:// <- that url00:38
Lotus907efiI meant the file does exist but not in that short path00:38
lifelessthats not a short path00:38
Lotus907efiinstead it is at:
lifelessits the correct path for the mirror URL you provided.00:39
lifelessLotus907efi: what mirroring tool are you using to make your local mirror ?00:39
Lotus907efiI don't know. I did not set it up00:39
Lotus907efithe repo expert for hpcloud did it00:39
Lotus907efialso known to you as jeb00:39
lifelessso to use DIB_DISTRIBUTION_MIRROR *at all*, you need to have a single unified mirror, not the partitioned thing you have.00:40
Lotus907efiyeah, I see that00:40
lifelessif jeb is using apt-mirror its a simple configuration setup00:40
lifelessthe apt format supports mixing all this easily00:40
Lotus907efiso I am going to dump using that and alter the devtest_overcloud.sh end user image build to include apt-sources00:41
Lotus907efiand see if that works00:41
Lotus907efijeb does not really "appreciate" when I suggest he do things differently  :)00:42
lifelessI'll be happy to supply a working apt-mirror configuration00:42
tchaypoIt's time for me to ask a stupid question. If we're planning to jettison the seed vm by making the undercloud self-aware and able to manage itself, why don't we also jettison the underclould and make the overcloud self-aware?00:43
Lotus907efiyeah, I can easily set that up on any of the other 15 - 20 systems I have00:43
tchaypoI suspect I've just revealed a bunch of things I don't understand yet00:43
lifelesstchaypo: several things00:44
Lotus907efi"customers' using systems in the overcloud should not know the undercloud is even there00:44
lifelesstchaypo: a) ^00:44
Lotus907efiin our cloud setup the sepearation of undercloud or operations from the overcloud is a security thing00:44
lifelesstchaypo: b) its nontrivial to mix hypervisors in one region00:45
*** EntropyWorks has quit IRC00:45
*** jcooley has quit IRC00:45
lifelesstchaypo: c) if the undercloud and overcloud are mixed, how do you do a test deploy ?00:45
lifelessof an overcloud change, I mean00:45
Lotus907efiis there a d)  ??? :)00:46
Lotus907efia, b, and c sound like enough reasons for me00:46
tchaypoc) - are you talking about doing a partial-rollout of the new overcloud config, or are we talking about one undercloud managing >1 overclouds, with one used for testing upgrades and one production cloud?00:46
tchaypothe partial-rollout is more in line with what I'd understood00:46
lifelesstchaypo: the latter00:50
lifelesstchaypo: final pre-prod check, deploy to 5 machines or something00:50
tchaypogreghaynes, StevenK - the doubletree appears to be sold out, according to CWT. hrmf.00:51
tchaypolifeless: that alone is a pretty convincing reason00:51
*** fandi has joined #tripleo00:51
tchaypoand then i search again and it's not sold out00:52
lifelessSpamapS: got a second ?00:52
tchaypolifeless: I'd like to make sure I undersand b - I think you're saying that we might have situations where the undercloud is using kvm but the overcloud is using vmware, or something along those lines?00:52
lifelessSpamapS: I have a merge.py brainfuk00:52
lifelesstchaypo: undercloud uses nova-bm or ironic00:53
lifelesstchaypo: overcloud uses kvm or vmware or xen00:53
*** yamahata has joined #tripleo00:55
* lifeless fires up the debugger00:55
tchaypoIF you're calling ironic a hypervisor I think you mean something different by it than what I had imagined. I think of ironic as a deployer. I think maybe the difference you mean is that nodes in the undercloud are deployed by ironic, regardless of whether they're baremetal or something else; nodes in the overcloud are deployed by poking at the kvm/vmware/xen API?00:56
Lotus907efiwait, overcould can use VMware???00:56
*** yamahata_ has quit IRC00:56
*** yamahata_ has joined #tripleo00:56
lifelessLotus907efi: sure, if you use the vmware driver for nova00:56
lifelessLotus907efi: we have no experience with that ourselves, upstream, since its not open source00:57
Lotus907efiI did not know thta existed00:57
Lotus907efihmmm. ok00:57
lifelesstchaypo: Ironic is effectively a hypervisor - sure it doesn't do much in the way of resource isolation00:58
mordredit does - the resources are isolated by air :)00:58
mordredold skool00:58
lifelesstchaypo: but, stop / start / put an image on a thing / configure network around it etc00:59
lifelesstchaypo: from *nova*'s perspective it is precisely a hypervisoir.00:59
*** weshay has quit IRC00:59
Lotus907efihmm, interesting way of looking at a provisioning tool01:00
Lotus907efibut true I guess01:00
lifelesstchaypo: the overcloud is deployed by asking nova to deploy images, which it does via the ironic 'hypervisor' driver01:01
lifelessor nova-bm01:01
lifelessor - you can obviously deploy these images to vms and bring up a fully virtualised openstack01:01
lifelessbut we lose some testing fidelity which is why we don't do that01:01
lifelessNobodyCam: template still isn't right - -----END RSA PRIVATE KEY-----' -P NeutronPublicInterface=eth0 -P IronicPassword=2e2cbb4ca564b7ba3fedd923afd8c2bfa283bc6b undercloud01:01
lifelessERROR: The Parameter (IronicPassword) was not defined in template.01:02
lifelessNobodyCam: trying to figure it out now01:02
openstackgerritlifeless proposed a change to openstack/tripleo-heat-templates: Fix Ironic templates  https://review.openstack.org/8241401:05
lifelessNobodyCam: fixed01:05
*** UtahDave has quit IRC01:14
tchaypookay, flights and hotel booked. Time for lunch.01:15
lifelessclarkb: remind me, what should a wheeel + source pip.conf look like ?01:16
Lotus907efijang has one on his wiki page01:17
Lotus907efi[global]# The following two settings for running pypi-mirrorallow-all-external = trueallow-unverified =        pysendfile        pytidylib # The following for using the mirror for your openstack-tools venv bootstrapindex-url =
openstackgerritlifeless proposed a change to openstack/diskimage-builder: Remove an excess cp of disk images.  https://review.openstack.org/8268301:17
Lotus907efijust pM'ed it to you01:18
Lotus907efithe URL01:18
* tchaypo notes that uber is active in atlanta01:23
Lotus907efiI think my user.qcow image was built this time with my repo mirror01:25
* Lotus907efi lived in Atlanta suburb for 10+ years01:26
*** rwsu has quit IRC01:26
StevenKLotus907efi: http://en.wikipedia.org/wiki/Uber_%28company%2901:26
Lotus907efihmm, interesting  I had never heard of that01:28
*** nosnos has joined #tripleo01:29
*** matsuhashi has quit IRC01:30
*** matsuhashi has joined #tripleo01:30
Lotus907efiso it is more expensive than a taxi but you can watch the car arrive on your smartphone01:30
tchaypoand between your phone and the driver's, it uses gps to log your exact route and calculate fares01:31
tchaypoand payment happens magically, charged to your linked T&E card, with an invoice emailed to you for ease of expense claiming later01:31
StevenKI may get stabbed for this, but "taxis for hipsters"01:32
* tchaypo goes to fetch skinny jeans01:32
lifelessNobodyCam: needs another layer01:35
Lotus907efiyeah, that is what I thought too :)01:35
Lotus907efi"taxis for hipsters" that is01:35
* NobodyCam reads scroll back01:37
lifelesspre_start_hook is also triggering an API call01:37
* tchaypo really goes for lunch01:37
*** rwsu has joined #tripleo01:42
lifelessNobodyCam: I've got a good feeling about this time01:47
*** yamahata_ has quit IRC01:48
*** CaptTofu has quit IRC01:49
*** yamahata_ has joined #tripleo01:50
NobodyCam:)heat template looks good... will run thru first thing in the am... oh question lifeless.. Is ironic needed in the overcloud?01:52
lifelessNobodyCam: nope01:52
StevenKlifeless: Did you look over https://review.openstack.org/#/c/82431/ ?02:03
lifelessStevenK: I did but it hadn't passed CI02:11
lifelesshas now so +202:11
*** CaptTofu has joined #tripleo02:22
lifelessNobodyCam: devananda: holy crap, undercloud came up. SHIP IT02:25
lifelessovercloud has failed to deploy to one of the nodes02:25
clarkblifeless: http://paste.openstack.org/show/74204/ for wheel + source pip.conf02:28
lifelessclarkb: 12.04 - which arch - i386 or amd64 ?02:30
clarkbwhatever the mirror was built on02:30
lifelesshmm, wheels do seem to encode arch02:30
clarkbso for us its all amd6402:31
clarkbif you build your own on i386 you will get i38602:31
*** mestery has joined #tripleo02:34
NobodyCamlifeless: did you see if the nodes got created in your undercloud ironic?02:34
lifelessNobodyCam: they did02:34
lifelessNobodyCam: I'm filing a bug now02:34
lifelessNobodyCam: HTTPInternalServerError' object has no attribute '__name__02:43
*** mestery has quit IRC02:46
*** mestery has joined #tripleo02:46
adam_glifeless, i was just debugging something similar02:49
*** fandi has quit IRC02:50
adam_ghttp://paste.ubuntu.com/7149391/ gave me some more insight02:51
lifelessNobodyCam: adam_g: https://bugs.launchpad.net/ironic/+bug/129706302:52
uvirtbotLaunchpad bug 1297063 in ironic "failed to deploy overcloud node" [Undecided,New]02:52
lifelessnow, lets see if nova delete works02:55
adam_glifeless, looks like a config error on the conductor side (assuming you're using pxe), make sure [conductor] has api_url  set to http://$host:6385. still a bug around exception handling thereon the driver side02:58
adam_gdelete should be working again if you're up to date /w https://review.openstack.org/#/c/82669/02:58
lifelessadam_g: ah no it merged after mine02:59
lifelessadam_g: no, the config error isn't an error03:00
lifelessadam_g: thats the bootstrap sorting itself out03:00
lifelessadam_g: you can see that the other two instances deployed ok03:01
* adam_g looks closer03:01
lifelessadam_g: it may be that there's a cached value and it re-reads after that, or something03:02
lifelessadam_g: you might like to reproduce, full instructions are invluded in the bug ;)03:02
adam_glifeless, yeah-- will take a look tomorrow. im about EOD.03:02
lifelessI'm now satisfied that the ironic patches are good03:03
lifelessso its a matter of nagging SpamapS slagle Ng GheRivero etc to review the tie and tht ones and perhaps land-with-failures the t-i one03:03
*** cwolferh has quit IRC03:03
*** CaptTofu has quit IRC03:09
*** matsuhashi has quit IRC03:12
*** cwolferh has joined #tripleo03:16
StevenKlifeless: I'm going to rebase no-more-backwards-compat on top of the --nodes patch03:20
StevenKNow that I've stopped fighting with flights and such03:20
*** nosnos has quit IRC03:29
*** matsuhashi has joined #tripleo03:30
*** vkozhukalov has joined #tripleo03:35
StevenKlifeless: And I see your single line has no love at all03:35
StevenKid=$(nova baremetal-node-create $(jq -r '"--pm_address=...03:35
*** matsuhashi has quit IRC03:36
lifelessStevenK: no love?03:42
StevenKlifeless: https://review.openstack.org/#/c/78467/11/scripts/register-nodes03:42
lifelessthats kinda faster03:43
lifeless76-neutron                                   10.71203:43
lifeless75-glance                                     6.10203:43
lifeless74-nova                                      10.82303:43
lifelessbut I reckon we can make that faster if we avoid all the re-entry03:43
lifeless(it looks like we call pip once per line in requirements...)03:44
openstackgerritlifeless proposed a change to openstack/diskimage-builder: Permit use of wheel mirrors in pypi element.  https://review.openstack.org/8270503:48
lifelessLotus907efi: jang: https://review.openstack.org/82705 may interest you03:49
* StevenK ponders an evil for and eval loop03:50
StevenKeval $var=$(jq ".$var" <<< $NODE_JSON)03:51
tchaypoIf any cores other than lifeless would care to take a look - https://review.openstack.org/#/c/80920/ is looking for another +203:51
StevenKtchaypo: We have other cores in this TZ?03:51
lifelesswe will if you guys review lots ;)03:51
tchaypohttps://review.openstack.org/#/c/82405/ is in the same state03:52
lifelessaim for 3+ reviews a day to get good coverage of changes03:52
tchaypoStevenK: I'd rather say that we have lots of people who work unusual hours *looks at ng*03:52
lifelessStevenK: I knew which review, I don't understand 'no love'03:52
StevenKlifeless: The two -1's WRT to "Please split that line up, my eyes are bleeding"03:53
lifelessStevenK: oh, folk have reviewed.c ool.03:53
lifelessStevenK:  \  might be your friend there.03:53
StevenKThey've both suggested seperate variables03:54
lifelessStevenK: what do you think ?03:54
tchaypospeaking of reviews - I'm finding it fairly common for me to give a +1 and then other people step in and object to something because of factors i hadn't considered or didn't understand. I'm finding it a good way to learn what our expected standards are, and I can't see any downside aside from getting a reputation for being easy to please..03:54
StevenKI'm just pondering eval and a for loop to set seven variables03:54
lifelesstchaypo: thats exactly it :)03:54
tchaypogood :)03:55
tchaypoi don't suppose you're core in infra?03:55
lifelessthey look for puppet changes specifically03:55
lifelesswhich isn't my forte03:55
StevenKHuh, I didn't know about adding my address to gitdm03:55
lifelessI am core on some things like pbr03:56
greghaynesStevenK: After the hard time you gave me for my use of eval!?03:59
StevenKgreghaynes: What do you think I'm currently giving myself?04:00
StevenKI *really* dislike eval04:00
greghaynesWonderful, ill only give your review a little bit of a hard time then :p04:01
StevenKWill you give it no hard time if it doesn't use eval?04:01
tchaypooh dear StevenK04:02
tchaypoI'm completly horrified04:02
greghaynesIts a possibility. I just like that youre succumbing to the eval usage04:03
StevenKgreghaynes, tchaypo: http://paste.openstack.org/show/74213/04:05
tchaypoto be clear, it's https://review.openstack.org/#/c/81691/ that has me completly horrified, not the use of eval04:05
tchaypouse of eval would have me completely horrified04:05
* tchaypo pokes uvirtbot04:05
*** killer_prince is now known as lazy_prince04:06
lifelessgreghaynes: here, have a '.04:06
openstackgerritGregory Haynes proposed a change to openstack/diskimage-builder: Enable wheel based pypi mirror  https://review.openstack.org/8270704:07
tchaypois it overly pedantic of me to complain that https://review.openstack.org/#/c/82431/1/scripts/setup-baremetal contains stylistic fixes (replacing ` ` with $( ) ) intermingled with substantive changes?04:09
StevenKYes, I also dislike ever using ``04:12
greghaynesclarkb: Your pypi config suggestion seems to work ^04:12
lifelessgreghaynes: uhm04:12
lifelessgreghaynes: https://review.openstack.org/8270504:12
*** nati_ueno has quit IRC04:13
lifelessits less minimal but more featureful04:13
* greghaynes abandons04:13
StevenKgreghaynes: I don't think you can call rewriting a script from bash to python as a 'drive-by' :-)04:14
lifelessStevenK: 15m ?04:14
StevenKlifeless: Hm?04:15
lifelessStevenK: my rewrite04:15
lifelessgreghaynes: also sorry, didn't realise you were poking at that04:16
greghaynesIts np, I like your change better anyways :)04:17
greghayneswoo more python04:17
*** matsuhashi has joined #tripleo04:23
greghayneslifeless: mind if I co-author some fixes for that one?04:24
greghaynesneeds a mkdir04:24
StevenKHuh, when did ip route replace start existing04:25
*** nosnos has joined #tripleo04:25
greghayneshrm, maybe thats not it... its def. failing for some reason though :/04:26
greghaynesah, because you still need a PYPI_MIRROR_URL set to something.. *facepalm*04:28
tchaypo"temporary failure in name resolution" *grumble grumble*04:35
tchaypowhere's that etherpad about all the env variables and how we want to stop doing that?04:38
greghaynesrationalise-ui IIRC04:39
*** nati_ueno has joined #tripleo04:40
*** lazy_prince2 has joined #tripleo04:41
*** lazy_prince2 is now known as killer_prince04:41
tchaypoyep, found it in my logs04:45
*** ramishra_ has joined #tripleo04:45
tchaypoand then dropped it on StevenK's review04:45
tchaypobut this raise another issue04:46
tchaypothis is config specific to one DIB element04:46
tchaypomayhap the directory layout for the config should be expanded to include per-element settings?04:46
SpamapSlifeless: back now04:48
tchaypoactually this probably fits within directories already listed there - the dib-pip-manifests directories04:48
tchaypoSpamapS: query as to core status04:49
SpamapSanswer for the win04:50
lifelessSpamapS: o/04:52
lifelessSpamapS: lots of reviews pending +A :)04:53
lifelessgreghaynes: it does?04:53
lifelessgreghaynes: cause, I include a mkdir04:53
SpamapSlifeless: ok, I'll do a few shortly.. just tying things up before bed04:53
lifelessSpamapS: cool04:53
tchaypoSpamapS: specifically https://review.openstack.org/#/c/80920/ and https://review.openstack.org/#/c/82405/ are the two i care most about04:54
lifelessSpamapS: backscroll over the last couple hours has a few refs04:54
tchaypoothers may have different priorities ;)04:54
StevenKlifeless: So I've added " via dev virbr0  src" but ping gives ICMP redirects and connections hang04:54
lifelessStevenK: eth1 on your seed is not up04:55
lifelessStevenK: at an educamated guess04:55
StevenKHm, it's not ubuntu/ubuntu ?04:56
lifelessStevenK: stack/stackl04:56
StevenKHm, you're right, and it's not configured via e/n/i04:57
*** yamahata has quit IRC04:58
lifelesscheck os-refresh-config.log04:58
*** yamahata has joined #tripleo04:58
StevenKIt was saved to disk, not shutdown and rebooted04:59
StevenK/var/log/os-apply-config.log mentions nothing about eth104:59
*** yamahata has quit IRC04:59
*** yamahata has joined #tripleo04:59
openstackgerritSteve Kowalik proposed a change to openstack/tripleo-incubator: Completely subsume POWER_MANAGER into testenv  https://review.openstack.org/8169105:01
*** eghobo has joined #tripleo05:01
lifelessStevenK: os-collect-config05:02
lifelessStevenK: not apply05:02
StevenKlifeless: collect-config is more interesting05:03
StevenKIt downed eth1 for some reason05:03
*** tzumainn has quit IRC05:03
StevenKLooks like 80-neutron-openvswitch-agent deleted from eth105:04
lifelessit shold05:04
lifelessit needs to move it to br-ctlplane05:04
*** vkozhukalov has quit IRC05:05
tchaypoI've finally understood why i sometimes see "dirname: invalid option -- 'b'05:05
tchaypoit's because things are looking for dirname $0, but $0 is set to -bash05:05
StevenKlifeless: br-ctlplane has no IP at all05:07
lifelesstchaypo: doh05:08
openstackgerritA change was merged to openstack/tripleo-incubator: Remove $(dirname $0) from devtest.sh docs.  https://review.openstack.org/8092005:10
*** CaptTofu has joined #tripleo05:10
killer_princeSpamapS: thanks for reviewing https://review.openstack.org/#/c/7928905:12
killer_princeSpamapS: but i have some doubts regarding downgrade issue that you have pointed out..05:13
killer_princeSpamapS: would it be possible for you to elaborate a bit..05:14
openstackgerritA change was merged to openstack/tripleo-incubator: Stop devtest if user was not in libvirtd group  https://review.openstack.org/8240505:14
*** matsuhashi has quit IRC05:14
*** vkozhukalov has joined #tripleo05:15
SpamapSkiller_prince: It is a corner case, no doubt, but it is not that hard to solve.05:15
*** CaptTofu has quit IRC05:15
killer_princeSpamapS: before i solve, i need to understand the problem.. can you pls help..05:16
SpamapSkiller_prince: basically by just looking for the current kernel version without any anchors, you match versions that have the string, plus more, in them.05:16
tchaypoSpamapS: and now you have lead me to discover that stipulate can mean "having stipules"05:17
SpamapSa small leaflike appendage to a leaf, typically borne in pairs at the base of the leaf stalk.05:17
SpamapStchaypo: actually reading the official definition of stipulate makes me think it is appropriate.05:18
*** matsuhashi has joined #tripleo05:18
killer_princek.. i get it now..... thanks..05:19
tchaypolooking at http://www.merriam-webster.com/dictionary/stipulate I think it's what's listed there as the transitive verb, wherease you were probably thinking of the intransitive verb05:19
tchaypoand now I need to look up what transitive and intransitive verbs are05:19
*** xuhaiwei has joined #tripleo05:19
StevenKtchaypo: Back to high school English with you?05:20
tchaypoI went to high school in QLD in 1994-1999. We learnt about verbs, nouns, and adjectives05:21
tchaypoeverything else i know about grammar I learnt from learning german05:21
xuhaiweiYour high school takes 5 years?05:22
SpamapStchaypo: same here.. except from learning Spanish and more recently Czech :)05:22
SpamapSI re-learned what an infinitive form was this past weekend :)05:22
tchaypoIn the main, my understanding of how to use english comes from having done a lot of reading; I have intuitions about what's normal and what isn't but usually don't know why05:22
StevenKxuhaiwei: 6 years05:22
SpamapSin the US we had to go to high school for 10 years05:23
tchaypoxuhaiwei: it was 5 in QLD when i went to school, 6 in NSW at the same time (when StevenK was going through it)05:23
xuhaiweisounds crazy, my high school is only 3 years05:23
SpamapSwe studied in a box in the middle of the road05:23
tchaypoQLD is just about to switch to 6, so they're going to have a double-intake year05:23
StevenKWe have Kindergarten, years 1 to 6, which is primay school. Years 7 to 12 are high school05:23
xuhaiweiwhat is QLD05:24
SpamapSUS goes K-5 "Elementary School", 9-12 "High School"05:24
StevenKSpamapS: You have this strange elementary, middle and high school thing.05:24
SpamapSwe don't acknowledge middle school05:24
StevenKxuhaiwei: Queensland, a state in Australia05:24
SpamapSit is a dark, terrible place05:24
tchaypoSpamapS: 9-12 seems like it's 6 years short of your claimed 10...05:25
StevenKSpamapS: We had a American do years 9-12 at my school because she was so stressed about the SATs05:25
xuhaiweiAustralians' mother language is English, isn't it?05:25
SpamapSwhere puberty bursts forth upon the unsuspecting students, gushing forth like a river of insanity, berthing boats populated iwth pure lust and whole barges full of awkwardness.05:25
clarkbI went to guam to take the SAT05:25
tchaypoxuhaiwei: yes, but we don't study it05:26
SpamapSclarkb: in soviet guam, SAT goes to you.05:26
tchaypoxuhaiwei: most of us don't study any other languages either, except for maybe a couple of years in primary school05:26
SpamapSYou don't study english directly?05:26
StevenKIt's more slow form torture, rather than studying.05:27
SpamapSMy god.. we do that right up until 11th grade when the "English" class really turns into "Literature"05:27
xuhaiweitchaypo: It's too lucky for you to study one language only05:27
SpamapSGrammar and spelling and composition.. tons of it.05:27
tchaypoWe had a class called "english" but it was all about analysing protest poetry and staging bits of shakespeare and so on05:27
StevenKtchaypo: Should have done Contemporary?05:27
SpamapStchaypo: yeah, that's the last 2 years in the US05:27
StevenKI skipped all that by doing the lowest English05:28
SpamapSIt's a bit of that in years 9/10 .. but 11 is where they annointed us with shakespearean confusion.05:28
StevenKInstead, we spent five months studying All Quiet on the Western Front05:28
SpamapSStevenK: do you fail if you speak? ;)05:28
tchaypoxuhaiwei: it leaves me embarrassed whenever I travel overseas - everywhere I go I'm surrounded by people I can only communicate with because they took the time to learn my language05:28
StevenKAnd then another five on Strictly Ball... no, can't remember. Too painful. :-P05:28
tchaypoStevenK: what's this "contemporary" thing? in QLD we just had "english"05:29
*** cwolferh has quit IRC05:29
SpamapSxuhaiwei: do you know the english word for a person who speaks two languages?05:29
clarkbtchaypo: it was great when I went to europe, the languages almost make sense05:30
StevenKtchaypo: Years 11 and 12 when I went through had four levels -- 2U Contemporary, 2U <Something that I've forgotten>, 2U Related and 3U English05:31
xuhaiweitchaypo: If in school you are forced to learn a language which you don't know whether will be used in the future, it's really upset05:31
killer_princeSpamapS: so basically i have to take care of special chars like . - etc which can exist in the kernel version.. right..05:31
xuhaiweiSpamapS: no05:31
tchaypoxuhaiwei: that's true05:31
*** cwolferh has joined #tripleo05:32
StevenKHm, rebooting my seed did not help, which saddens me.05:32
SpamapSxuhaiwei: "bilingual"05:32
SpamapSxuhaiwei: and 3 languages is "trilingual"05:32
SpamapSxuhaiwei: but 1 language... care to guess?05:32
SpamapSkiller_prince: no05:32
SpamapSkiller_prince: you have to make sure you find the "end" of the kernel version05:33
StevenKSpamapS: "American" ? :-P05:33
SpamapSStevenK: well done :)05:33
xuhaiweiSpamapS: lingual?05:33
SpamapSxuhaiwei: ^^05:33
SpamapSStevenK has met a few of us.05:33
xuhaiweiIs that a trap?05:33
SpamapSjust a joke :)05:36
tchaypoit's not fair though - most australians are also monolingual05:38
StevenKtchaypo: But we also don't travel everywhere demanding everyone speak English.05:38
tchaypoand the once who do have another language usually have it because their parents/grandparents were immigrants and their mother tongue is still used at home05:38
StevenKWe just travel everywhere and work in bars.05:38
*** vkozhukalov has quit IRC05:39
tchaypoStevenK: am i correct in thinking that https://review.openstack.org/#/c/80949/2/scripts/setup-clienttools would interfere with the thing you did that rewrites ~/.pydistutils.cfg ?05:40
killer_princeSpamapS: sorry but i am still confused. I extract the line with kernel and then I am going token by token using a for loop. and then i am extract the kernel version by removing /boot/vmlinux- which means what is left is just kernel version.. is there anything that i missed..05:42
StevenKtchaypo: I wrote something that rewrites what?05:44
tchaypohrm. I could swear i just read something you wrote where a shell script was replaced with some python that munges ~/.pydistutils.cfg05:47
*** nati_ueno has quit IRC05:47
tchayponope, I'm thinking of https://review.openstack.org/#/c/82705/1/elements/pypi/pre-install.d/00-configure-pypi-mirror from lifeless05:48
*** e0ne has joined #tripleo05:49
*** e0ne has quit IRC05:49
StevenKtchaypo: I'm easily confused with lifeless? :-P05:50
*** nati_ueno has joined #tripleo05:51
tchayponow I'm double-confused; lifeless gave https://review.openstack.org/#/c/80949/2 a -2 because (if I'm reading his comment correctly) it tramples on ~/.pydistutils.cfg; then he wrote 82705, which does that and more.05:52
tchaypoI assume I'm missing some context05:52
lifelesstchaypo: 82705 runs within the chrooted image build05:59
lifelesstchaypo: 80949 affects the users actual home dir05:59
tchaypolifeless: oh. that's quite some context.06:00
lifelesstchaypo: looking at the code base in question should supply that, in general06:01
tchaypoyeah. i know 82705 was a dib element. that should have been a clue.06:02
*** ramishra_ has quit IRC06:06
tchaypoand on that note, I'm out for the day.06:11
*** andreaf has joined #tripleo06:14
*** mestery has quit IRC06:18
SpamapSkiller_prince: that is all fine, but then in your sed statement you don't anchor that kernel version, and you replace it in the entire file.06:22
SpamapSkiller_prince: so s/${KERNEL_VERSION}[right here you need something to let sed know where the end of the kernel version is]/06:23
SpamapSkiller_prince: so if you expect there to always be a space, or end of line, or ??? there (I'm not really sure what you're aiming at and not super familiar with teh grub.conf details.) then that is what you need to use as an anchor06:25
SpamapSkiller_prince: without that, you will possibly run this:  s/3.11.0-11/3.11.0-1/ and would also change _existing_ instances of 3.11.0-1 to 3.11.0-11 .. which is entirely possible if the image has multiple kernels.06:28
SpamapShrm it is late, I messed up that scenario06:29
SpamapSyou'd change instances of 3.11.0-11.1 to 3.11.0-1.0 which may not exist.06:29
SpamapSkiller_prince: anyway, anchor your search string and all is good.06:30
* SpamapS goes to bed06:30
killer_princeSpamapS: thanks now i get it... i will fix it now..06:32
*** eghobo has quit IRC06:46
xuhaiweiI am starting devtest now, and following this instructions http://docs.openstack.org/developer/tripleo-incubator/devtest_setup.html, I have a question, do I have to create a seed vm?06:47
xuhaiweiI mean I have had the nova-baremetal worked, and use it to deploy a baremetal node successfully, but only by hand06:48
xuhaiweiI did that under the instruction of baremetal wiki06:49
lifelessxuhaiwei: do you have heat in your nova-baremetal setup ?06:49
xuhaiweinot yet06:49
lifelessto deploy a cloud using heat you need neutron nova keystone glance heat all operational; if you don't have that, the seed is the fastest way to get it06:50
xuhaiweithe seed is a openstack on the openstack?06:51
lifelessthe seed is a self contained nova baremetal06:51
lifelesswhich we use to deploy the undercloud06:51
*** ramishra has joined #tripleo06:51
lifelessusing openstack on openstack.06:52
lifelessAnd then we use the undercloud to deploy the overcloud, same way.06:52
*** ramishra_ has joined #tripleo06:53
xuhaiweiI saw the word 'seed vm', so I related it to a VM06:53
lifelessit is a vm06:54
*** ramishra has quit IRC06:54
lifelesssince we have no way to deploy it except by hand06:54
lifelesswe use direct virsh calls to deploy it06:54
lifelessand after that we can use the Nova API06:54
lifelessI believe this is all explained in the docs, but if not we should do better.06:54
*** ramishra_ has quit IRC06:55
xuhaiweiSo, the first step to build the seed is working on a clean Ubuntu(if i use ubuntu)?06:55
lifelessor fedora or whatever.06:56
lifelesswe try to be minimally invasive06:56
lifelessbut to try 'devtest' you'll want about 16GB of ram06:56
lifelesssince devtest is a demonstration, without the use of baremetal06:56
lifelessI mean it uses nova-baremetal06:57
lifelessbut you don't need physical machines to walk through the process06:57
xuhaiweiI think I understand it now, thank you06:57
lsmola_good morning tripleo07:01
*** lsmola_ has quit IRC07:02
*** lsmola_ has joined #tripleo07:02
*** vkozhukalov has joined #tripleo07:03
*** eghobo has joined #tripleo07:05
*** bogdando has quit IRC07:05
*** andreaf has quit IRC07:08
*** CaptTofu has joined #tripleo07:11
*** ramishra has joined #tripleo07:13
*** CaptTofu has quit IRC07:16
*** jtomasek has joined #tripleo07:16
GheRiveromorning all07:28
*** nati_ueno has quit IRC07:37
*** xuhaiwei has quit IRC07:37
*** rdopieralski has joined #tripleo07:40
*** ifarkas has joined #tripleo07:48
*** eghobo has quit IRC07:54
*** rlandy has joined #tripleo07:55
openstackgerritLadislav Smola proposed a change to openstack/tripleo-heat-templates: WIP heat stack-update quick fix  https://review.openstack.org/8247907:55
*** bogdando has joined #tripleo08:03
*** e0ne has joined #tripleo08:10
*** jprovazn has joined #tripleo08:10
lifelessnova bm is hitting me hard today08:12
lifeless| fault                                | {"message": "No valid host was found. ", "code": 500, "details": "  File \"/opt/stack/venvs/nova/local/lib/python2.7/site-packages/nova/scheduler/filter_scheduler.py\", line 108, in schedule_run_instance |08:12
*** fandi has joined #tripleo08:15
rpodolyakamorning all08:21
*** xuhaiwei has joined #tripleo08:23
lifelesssomething has made nova-bm horribly flaky ;(08:28
*** killer_prince has quit IRC08:30
*** mrunge has joined #tripleo08:32
*** d0ugal has joined #tripleo08:34
*** killer_prince has joined #tripleo08:37
openstackgerritGhe Rivero proposed a change to openstack/tripleo-incubator: Add elements to user image via USER_DIB_EXTRA_ARGS  https://review.openstack.org/8273908:40
lifelessGheRivero: https://review.openstack.org/#/c/81944/ - any chance of a +A ?08:41
lifeless[well a +2 + +A :P]08:42
GheRiveroon it08:43
openstackgerritGhe Rivero proposed a change to openstack/tripleo-incubator: Add elements to user image via USER_DIB_EXTRA_ARGS  https://review.openstack.org/8273908:43
openstackgerritA change was merged to openstack/tripleo-incubator: Remove non-bootstrap code from init-keystone.  https://review.openstack.org/8194408:46
*** giulivo has joined #tripleo08:46
openstackgerritlifeless proposed a change to openstack/tripleo-incubator: Update documentation.  https://review.openstack.org/8274108:47
*** matsuhashi has quit IRC08:49
*** jistr has joined #tripleo08:49
lifeless2014-03-25 08:26:33.396 4757 TRACE nova.compute.manager [instance: 9939a246-60fd-47ff-b53b-ac4d7d2dee6b] RuntimeError: Second simultaneous read on fileno 13 detected.  Unless you really know what you're doing, make sure that only one greenthread can read any particular socket.  Consider using a pools.Pool. If you do know what you're doing and want to disable this error, call eventlet.debug.hub_prevent_multiple_readers(False)08:49
*** athomas has joined #tripleo08:50
lifeless'yay' https://bugs.launchpad.net/nova/+bug/122947508:52
uvirtbotLaunchpad bug 1229475 in python-neutronclient "terminate_instance(): RuntimeError: Second simultaneous read on fileno 16 detected" [High,Fix committed]08:52
*** pblaho has joined #tripleo08:53
*** andreaf has joined #tripleo08:53
*** jistr is now known as jistr|training08:54
*** matsuhashi has joined #tripleo08:56
*** viktors has joined #tripleo08:57
*** jcoufal has joined #tripleo08:57
*** gcha has joined #tripleo08:59
openstackgerritOm Kumar proposed a change to openstack/tripleo-image-elements: Adds local storage Boot support in case PXE boot fails.  https://review.openstack.org/7928909:00
*** hashar has joined #tripleo09:02
lifelessGheRivero: https://review.openstack.org/#/c/82705/ may interest you too09:10
*** CaptTofu has joined #tripleo09:12
*** matsuhashi has quit IRC09:14
*** derekh has joined #tripleo09:15
*** CaptTofu has quit IRC09:16
*** lucasagomes has joined #tripleo09:17
*** matsuhashi has joined #tripleo09:18
openstackgerritDerek Higgins proposed a change to openstack/tripleo-incubator: Set selinux default to permissive on Fedora  https://review.openstack.org/8182609:21
*** tzumainn has joined #tripleo09:30
openstackgerritJan Provaznik proposed a change to openstack/tripleo-image-elements: Adds mysql-dev and mariadb-dev elements  https://review.openstack.org/7988609:30
openstackgerritJan Provaznik proposed a change to openstack/tripleo-image-elements: Adds mariadb element  https://review.openstack.org/7988409:30
openstackgerritJan Provaznik proposed a change to openstack/tripleo-image-elements: Adds mysql-common element  https://review.openstack.org/7988509:30
lifelessGheRivero: I've put some notes at the bottom of https://bugs.launchpad.net/tripleo/+bug/129573209:34
uvirtbotLaunchpad bug 1295732 in tripleo "tripleo devtest overcloud startup takes too long " [High,Triaged]09:34
GheRiverolifeless: thanks lifeless. Will talk with Ng to work on that09:35
Nglifeless: I had weird skype issues, but I was partially on the call just now. I heard you talking about virtio, but I couldn't tell if you were saying that someone is already looking at that, or if nobody is09:36
lifelessGheRivero: np; I'll be poking more tomorrow. https://review.openstack.org/#/c/82683/ would be a cheap win of a few seconds... if it worked09:36
lifelessNg: someone should09:36
lifelessNg: I think its a key factor - I see 100% CPU utilisation copying bulk data over the network.09:36
GheRiverolifeless: Ng: I have spin a tripleo with virtio now, just waiting to finish09:37
NgI think we can also use virtio for the qcow2 image, which might help a bit09:37
Ngbut we don't support that atm in configure-vm. I'm almost done teaching it to do that09:38
lifelessman we really need that pypi mirror in place09:38
lifelessNg: virtio for disk? We can't until we fix nova.09:38
lifelessNg: or find some other way to ensure that ephemeral flavors work.09:38
openstackgerritlifeless proposed a change to openstack/tripleo-incubator: Default to 30G of disk per bm VM.  https://review.openstack.org/8275309:39
*** nosnos has quit IRC09:39
Nglifeless: hrm, why would they not work?09:41
lifelessNg: the ephemeral mapping from nova has sda hard coded in it09:42
lifelessNg: virtio disk is vda09:42
Ngugh :)09:43
lifelessNg: its a fairly straight forward patch to nova, but we need to do that, get a FF, get it in, then we can use virtio disks.09:43
*** bauzas has joined #tripleo09:43
lifelessNg: if virtio disks will make a difference, lets do it. But being able to do upgrades is probably more important than speed until then09:43
GheRiverois there the same problem in ironic?09:44
lifelessGheRivero: until we fix it yes :)09:46
*** killer_prince has quit IRC09:49
*** andreaf has quit IRC09:49
*** mrunge has quit IRC09:50
rcarrillocruzguys, do you get IndexError: list index out of range when running tox on tripleo?09:52
rcarrillocruzgetting that with tox 1.6 and 1.709:53
lifelessnew tox is broken09:53
lifelessthey forgot about unix compat09:53
rcarrillocruzlifeless: which tox ver do you use that is known to work?09:53
lifelessI believe 1.5 is09:54
rcarrillocruz1.6 not working for me, and it's not the latest09:54
rcarrillocruzk, let me try09:54
*** thouveng has joined #tripleo09:54
rcarrillocruzvagrant@vagrant-ubuntu-precise-64:~/tripleo-incubator$ tox09:56
rcarrillocruzERROR: tox version is 1.5, required is at least 1.609:56
rcarrillocruzi'll keep digging09:56
lifelessdunno then09:58
lifelesstox is run on most things on commit09:58
lifelessI'm not sure there is any tox stuff to run on the incubator itself09:58
rcarrillocruzk, i usually run it to make sure my stuff is ok (pep8, syntax and whatever is tested), but I guess can run plain pep8. tbh, i have done mostly openstack-infra stuff till now, so that may be why i didn't hit it ....10:03
*** mrunge has joined #tripleo10:04
*** mrunge has quit IRC10:04
*** mrunge has joined #tripleo10:04
*** morganfainberg is now known as morganfainberg_Z10:05
*** matsuhashi has quit IRC10:08
*** matsuhashi has joined #tripleo10:10
openstackgerritRicardo Carrillo Cruz proposed a change to openstack/tripleo-incubator: Add trap to clean up openstack-tools folder on failure Closes-Bug: 1226155  https://review.openstack.org/8276310:15
*** e0ne_ has joined #tripleo10:16
*** jp_at_hp has joined #tripleo10:16
openstackgerritA change was merged to openstack/tripleo-image-elements: Update elements to use os-svc-daemon -i.  https://review.openstack.org/7696310:17
openstackgerritRicardo Carrillo Cruz proposed a change to openstack/tripleo-incubator: Add trap to clean up openstack-tools folder on failure  https://review.openstack.org/8276310:17
*** killer_prince has joined #tripleo10:18
*** akrivoka has joined #tripleo10:18
*** e0ne has quit IRC10:19
openstackgerritDerek Higgins proposed a change to openstack/tripleo-incubator: Set selinux default to permissive on Fedora  https://review.openstack.org/8182610:19
*** CaptTofu has joined #tripleo10:25
*** killer_prince has quit IRC10:29
openstackgerritGhe Rivero proposed a change to openstack/tripleo-incubator: Allow to pass uncompress option to boot-seed-vm  https://review.openstack.org/8277210:30
howleytlifeless: I'm not following your concern over vhosts in rabbitmq. You said "I mean that neutron and cinder and nova will be using a common RPC service. Having vhosts will be problematic there." in https://etherpad.openstack.org/p/tripleo-vhost-for-rabbitmq-element10:35
*** athomas has quit IRC10:36
lifelesshowleyt: maybe I'm misunderstanding something10:39
lifelesshowleyt: the goal of vhosts is to partition the different services up, no ?10:39
howleytlifeless: it just partitions the resources on the rabbitmq server. So I can specify that the cinder service (and clients) only has permissions on queues defined in cinder vhost, for example.10:40
howleytbut cinder clients or service would not have access to other client's queues, unless we had specific ACL for that.10:41
lifelesshowleyt: right, so how do you configure it for say ceilometer which needs to be talked to by all the services10:42
lifeless(AIUI anyhow; if it doesn't, then gantt which I do understand can be used as an example)10:42
howleytlifeless: so I guess if ceilometer has a bunch of queues on a ceilometer vhost, all services would need permissions that allow them to publish messages to queues on the ceilometer vhost.10:44
*** xuhaiwei has quit IRC10:45
lifelesshowleyt: how do we configure nova then to talk to two vhosts?10:45
howleytlifeless: not so sure about that.10:46
lifelesshowleyt: so what I've done in the past to secure these things is use the rabbit regex based acls10:46
*** weshay has joined #tripleo10:47
lifelesshowleyt: would that meet your needs? I mean - I'm happy for a vhost config thing to land, I'm just not sure that we can use it :)10:47
*** andreaf has joined #tripleo10:47
*** killer_prince has joined #tripleo10:50
lifelessGheRivero: Ng: https://review.openstack.org/#/c/82705/10:50
howleytlifeless: I understand what you're saying. I'm not actually familiar with the ceilometer setup or how it uses rabbitmq.10:53
openstackgerritA change was merged to openstack/diskimage-builder: Permit use of wheel mirrors in pypi element.  https://review.openstack.org/8270510:54
*** ruagair has joined #tripleo10:54
lifelesshowleyt: it may be that it can work with a vhost10:55
lifelesshowleyt: haven't looked closely enough10:55
openstackgerritMichael Kerrin proposed a change to openstack/tripleo-incubator: WIP - configure a node ephemeral disk  https://review.openstack.org/8277611:04
*** hashar has quit IRC11:08
openstackgerritJames Slagle proposed a change to openstack/tripleo-image-elements: Add fedora-rdo-icehouse element  https://review.openstack.org/8065811:10
*** ramishra has quit IRC11:11
*** akrivoka has quit IRC11:12
openstackgerritJames Slagle proposed a change to openstack/tripleo-image-elements: Add fedora-rdo-icehouse element  https://review.openstack.org/8065811:13
lifelessslagle: either up up way to late, or you're up WWWAAAYYYYY to late11:16
slaglei'm up early here :)11:17
*** hashar has joined #tripleo11:17
slaglei am looking at snow. which i really thought was over for the season11:17
*** CaptTofu has quit IRC11:17
*** ramishra has joined #tripleo11:19
howleytlifeless: thanks, I'll take a look into it.11:19
*** CaptTofu has joined #tripleo11:21
*** athomas has joined #tripleo11:22
lifeless2014-03-25 11:06:17.636 | /opt/stack/new//diskimage-builder/bin/ramdisk-image-create: line 190: remove_image: command not found11:25
lifeless2014-03-25 11:06:17.773 | ERROR: gate_hook failed11:25
lifelessfrom http://logs.openstack.org/83/82683/1/check-tripleo/check-tripleo-seed-precise/eaf118f/console.html11:25
lifelessfrom https://review.openstack.org/#/c/82683/11:25
lifelessNg: since you did the ramdisk stuff I'm thinking you might be best placed to trivially fix my patch :)11:25
lifelessNg: shaving a few more seconds of waste....11:26
*** athomas_ has joined #tripleo11:26
lifelessNg: / slagle: - https://review.openstack.org/#/c/82753/ very useful for -u images :)11:26
lifelessnight all!11:26
*** athomas has quit IRC11:29
openstackgerritlifeless proposed a change to openstack/tripleo-image-elements: Fixup Ironic support.  https://review.openstack.org/8162711:31
openstackgerritlifeless proposed a change to openstack/tripleo-image-elements: Permit overriding the nova compute manager.  https://review.openstack.org/8267711:31
*** e0ne_ has quit IRC11:33
*** e0ne has joined #tripleo11:33
*** e0ne has quit IRC11:38
*** jistr|training has quit IRC11:40
*** hashar has quit IRC11:40
*** athomas has joined #tripleo11:40
*** akrivoka has joined #tripleo11:40
*** athomas_ has quit IRC11:41
*** jistr has joined #tripleo11:41
*** jistr is now known as jistr|training11:41
*** ifarkas has quit IRC11:43
*** fandi has quit IRC11:47
*** ifarkas_ has joined #tripleo11:48
*** ifarkas_ has quit IRC11:50
*** ifarkas has joined #tripleo11:51
*** sseago has quit IRC11:54
*** morazi has joined #tripleo11:58
*** bcrochet has quit IRC11:59
*** bcrochet has joined #tripleo12:01
*** pbelanyi has joined #tripleo12:06
*** e0ne has joined #tripleo12:12
*** e0ne has quit IRC12:13
*** julim has joined #tripleo12:13
*** e0ne has joined #tripleo12:14
*** matsuhashi has quit IRC12:16
*** e0ne has quit IRC12:18
*** matsuhashi has joined #tripleo12:23
*** yamahata has quit IRC12:27
*** CaptTofu has quit IRC12:45
*** lblanchard has joined #tripleo12:46
*** hashar has joined #tripleo12:46
*** lucasagomes is now known as lucas-hungry12:51
* slagle thinks jenkins ran out of coffee13:02
*** noslzzp has joined #tripleo13:02
*** matsuhashi has quit IRC13:03
killer_princeslagle: morning..13:03
*** dprince has joined #tripleo13:03
killer_princeslagle: i have updated https://review.openstack.org/#/c/79873/ with your comments. can you pls review it again..13:04
slaglekiller_prince: sure, will take a look13:04
killer_princeslagle: thanks in advance..13:05
*** markmc has joined #tripleo13:06
*** gcha has quit IRC13:13
*** spzala has joined #tripleo13:16
*** lazy_prince has quit IRC13:22
*** ChanServ changes topic to "the gerrit event stream is currently hung, blocking all testing. troubleshooting is in progress (next update at 14:00 utc)"13:22
*** lazy_prince has joined #tripleo13:24
*** matty_dubs|gone is now known as matty_dubs13:24
openstackgerritJon-Paul Sullivan proposed a change to openstack/tripleo-incubator: Add the select-cloud script  https://review.openstack.org/8280713:26
*** mestery has joined #tripleo13:28
*** mestery has quit IRC13:28
*** mestery has joined #tripleo13:29
openstackgerritLuigi Toscano proposed a change to openstack/diskimage-builder: Add console kernel parameters to extlinux configuration  https://review.openstack.org/8282213:30
*** ChanServ changes topic to "tripleo-cd running preserve-ephemeral WIP patches and https://review.openstack.org/#/c/62042/ | Using OpenStack to deploy OpenStack;meetings Tuesday 1900 UTC in #openstack-meeting-alt"13:30
*** ramishra has quit IRC13:30
*** lucas-hungry is now known as lucasagomes13:33
*** gcha has joined #tripleo13:34
*** killer_prince has quit IRC13:35
*** lazy_prince is now known as killer_prince13:35
openstackgerritJan Provaznik proposed a change to openstack/diskimage-builder: Adds "element-provides" support to element dependencies  https://review.openstack.org/7942113:38
*** yamahata has joined #tripleo13:38
*** nati_ueno has joined #tripleo13:40
geerdestzuul broken?  https://review.openstack.org/#/c/82807/13:42
*** hewbrocca has joined #tripleo13:43
*** fandi has joined #tripleo13:44
*** nati_ueno has quit IRC13:59
*** ramishra has joined #tripleo14:00
*** ramishra has quit IRC14:02
openstackgerritA change was merged to openstack/tripleo-incubator: Set selinux default to permissive on Fedora  https://review.openstack.org/8182614:03
*** ramishra has joined #tripleo14:04
*** hashar_ has joined #tripleo14:04
*** hashar has quit IRC14:05
openstackgerritA change was merged to openstack/tripleo-incubator: Add backports repo for jq package  https://review.openstack.org/8246314:21
*** mestery has quit IRC14:24
*** CaptTofu has joined #tripleo14:27
*** mestery has joined #tripleo14:33
openstackgerritJan Provaznik proposed a change to openstack/diskimage-builder: Adds "element-provides" support to element dependencies  https://review.openstack.org/7942114:36
*** dprince has quit IRC14:40
openstackgerrittom-howley proposed a change to openstack/tripleo-heat-templates: Add rabbitmq keepalive and keepalive parameters to overcloud template  https://review.openstack.org/8193414:40
*** dprince has joined #tripleo14:53
*** CaptTofu has quit IRC14:54
openstackgerritJon-Paul Sullivan proposed a change to openstack/tripleo-incubator: Add the select-cloud script  https://review.openstack.org/8280714:56
openstackgerritA change was merged to openstack/tripleo-image-elements: Glance: Add configuration hierarchy  https://review.openstack.org/7822614:59
*** sseago has joined #tripleo15:07
*** spzala has quit IRC15:10
devanandadprince: ping re https://review.openstack.org/8266815:11
dprincedevananda: hi, whats up?15:11
devanandadprince: hi! so I posted a different implementation of your patch15:12
devanandadprince: would like your input on it15:12
dprincedevananda: ah, so I looked at it an left feedback on my patch since that is where the thread was.15:12
dprincedevananda: the gist is I'm fine with either. I see yours as more of a refactoring of the code that was already there, and mine as just a bug fix.15:13
*** spzala has joined #tripleo15:13
devanandadprince: ah, gotcha. too many windows open, didn't see your reply15:14
dprincedevananda: noting that I do like the fact that your refactoring brings the ipmitool driver closer in line with the pygmi implementation15:14
*** spzala has quit IRC15:17
*** e0ne has joined #tripleo15:19
openstackgerritAna Krivokapic proposed a change to openstack/tuskar-ui: Fix checkboxes on the nodes page  https://review.openstack.org/8256815:21
*** noslzzp has quit IRC15:24
*** noslzzp has joined #tripleo15:24
*** mestery has quit IRC15:24
*** hashar has joined #tripleo15:33
*** hashar_ has quit IRC15:36
*** CaptTofu has joined #tripleo15:37
*** hashar_ has joined #tripleo15:38
dprinceslagle: would be nice to seal the deal on the Fedora em1/em2 stuff this week. https://review.openstack.org/#/c/82093/ and https://review.openstack.org/#/c/82144 should do it I think.15:40
*** hashar has quit IRC15:41
*** akrivoka has quit IRC15:42
*** akrivoka has joined #tripleo15:43
*** CaptTofu has quit IRC15:44
*** hashar_ has quit IRC15:49
openstackgerritA change was merged to openstack/diskimage-builder: baremetal: correct the path of ifcfg-eth0  https://review.openstack.org/8209315:50
*** eghobo has joined #tripleo15:52
*** eghobo has quit IRC15:52
*** eghobo has joined #tripleo15:53
openstackgerritA change was merged to openstack/tripleo-incubator: Fedora uses em1 for undercloud/overcloud  https://review.openstack.org/8214415:55
slagledprince: ^^15:56
*** killer_prince is now known as lazy_prince15:57
dprinceslagle: thanks, glad to finally have that off the list15:58
*** mestery has joined #tripleo15:59
*** rdopieralski has quit IRC16:00
*** matty_dubs is now known as matty_dubs|lunch16:02
*** fandi has quit IRC16:06
*** fandi has joined #tripleo16:08
*** mestery has quit IRC16:09
*** thouveng has quit IRC16:09
*** mestery has joined #tripleo16:10
*** gcha has quit IRC16:10
openstackgerritDerek Higgins proposed a change to openstack-infra/tripleo-ci: Expose more testenv parameters  https://review.openstack.org/8285516:10
*** viktors has quit IRC16:13
*** nati_ueno has joined #tripleo16:17
*** hashar has joined #tripleo16:20
SpamapSdprince: https://review.openstack.org/#/c/82144/1/scripts/devtest_overcloud.sh ... Would have been nice to test it on Ubuntu .. you _do_ have access to quite a few baremetal Ubuntu machines.16:21
dprinceSpamapS: I did test it :)16:21
SpamapSdprince: I'm pretty sure that we'll end up with em1 on Ubuntu.16:21
SpamapSunless baremetal was merged without installing biosdevname on Ubuntu16:21
dprinceSpamapS: It wasn't clear to me why it wasn't working though.16:22
*** newell has joined #tripleo16:22
*** ifarkas has quit IRC16:22
dprinceSpamapS: Based on feedback during the reviews I did my best to make baremetal support Ubuntu. And I tested it.16:22
SpamapSdprince: so Ubuntu has biosdevname, but still ends up with ethX ?16:22
dprinceSpamapS: Yes. I even hacked in the udev rules file to make it try it.16:23
*** vkozhukalov has quit IRC16:23
dprinceSpamapS: If I recall one of the udev rules files was guarding against using it as a default for some reason.16:23
SpamapSdprince: ah.. worth asking in #ubuntu-devel ... I'll do that16:24
dprinceSpamapS: gotta say. I did try it though. Perhaps I missed something stupid.16:27
SpamapSdprince: no, I think Ubuntu may be playing it safe. :p16:27
*** andreaf has quit IRC16:27
*** nati_ueno has quit IRC16:30
SpamapSdprince: Not really sure why though.. I'm taking a look at the udev rules now16:30
lsmola_SpamapS: hello16:30
SpamapSdprince: I swear I remember it being enabled right after 12.04 was released16:30
dprinceSpamapS: At first I tried manually appending the 'biosdevname' option (or whatever that udev rule expected).16:31
lsmola_SpamapS: any idea why this is not called during stack-update https://github.com/openstack/tripleo-image-elements/blob/master/elements/os-refresh-config/os-refresh-config/post-configure.d/99-refresh-completed#L1016:31
dprinceSpamapS: That seemed like the mechanism to use. Then that didn't work and I just freakin hackup up the udev rule, rebooted, and still no dice16:31
SpamapSdprince: I've seen 13.10 servers with emX ... so it may also be a hardware thing16:32
SpamapSdprince: as in, some of them do it, some don't16:32
dprinceSpamapS: well, technically I think only embedded NICs show up as em1/em216:32
SpamapSDunno, I'm building an image now and I'll test it.16:32
SpamapSlsmola_: that should definitely run16:32
dprinceSpamapS: em -> embedded16:32
SpamapSlsmola_: assuming completion-handle is set :)16:32
dprinceSpamapS: in my case this same box worked great for Fedora (and does have an embedded NIC on it)16:33
SpamapSdprince: right, so in the systems I've seen, eth0 and eth1 became em0 and em1, and then eth2 became em016:33
SpamapSdprince: with our mellanox cards for instance, I believe they represent themselves as embedded because the bios recognizes them16:33
SpamapSso then it goes from eth2 -> em216:34
dprinceSpamapS: Right. Other NICs get stable-names too BTW but sometimes they are funky. Like I have a couple USB NICs that show up as enp0s20u4. The important thing is it is consistent, and predictable... (unlike eth0, eth1...)16:35
SpamapSdprince: right.. so we want that .. not sure why we're not getting that16:36
SpamapSdprince: anyway, I've booted a box.. have a train to catch, will check back in from the rails. :)16:38
*** UtahDave has joined #tripleo16:39
*** bauzas has quit IRC16:40
openstackgerrittom-howley proposed a change to openstack/tripleo-image-elements: Add keepalive option to rabbitmq config  https://review.openstack.org/8192016:42
openstackgerritStuart McLaren proposed a change to openstack/tripleo-image-elements: Glance: Change seed json to use config hierarchy  https://review.openstack.org/7827116:46
lsmola_SpamapS: hmm my handle doesn't receive this signal, but id does with stack-create16:49
lsmola_SpamapS: I wonder why :-)16:50
*** jpeeler has joined #tripleo16:52
*** jpeeler has joined #tripleo16:52
lsmola_SpamapS: hmm I wonder if I should just hack this for now, since we should delete these handles anyway, with HOTSoftwareConfig16:53
lsmola_SpamapS: I've heard HOTSoftwareConfig is already handling that, right?16:54
*** jistr|training has quit IRC16:58
lsmola_SpamapS: so when update, it receives the signal only from the instances that were already created, but not from the new ones, weird :-o16:59
*** jcoufal has quit IRC17:00
openstackgerritDerek Higgins proposed a change to openstack/tripleo-image-elements: Add testenv image building to deploy-testenv  https://review.openstack.org/8168617:01
*** UtahDave has left #tripleo17:01
*** nati_ueno has joined #tripleo17:02
*** matty_dubs|lunch is now known as matty_dubs17:02
*** ifarkas has joined #tripleo17:03
*** morganfainberg_Z is now known as morganfainberg17:04
openstackgerritDerek Higgins proposed a change to openstack/tripleo-image-elements: Add testenv image building to deploy-testenv  https://review.openstack.org/8168617:07
openstackgerritDerek Higgins proposed a change to openstack/tripleo-image-elements: Add support to deploy-testenv for multiple environments  https://review.openstack.org/8168617:10
*** ramishra has quit IRC17:11
*** saurabhs has joined #tripleo17:12
*** ramishra has joined #tripleo17:13
*** saurabhs1 has joined #tripleo17:17
saurabhs1SpamapS : I split the big trove review in 4 different reviews, just making sure that is what you wanted me to do?17:17
*** saurabhs has quit IRC17:19
*** ifarkas_ has joined #tripleo17:20
openstackgerritRicardo Carrillo Cruz proposed a change to openstack/tripleo-incubator: Add trap to clean up openstack-tools folder on failure  https://review.openstack.org/8276317:24
openstackgerritChris Jones proposed a change to openstack/diskimage-builder: Remove an excess cp of disk images.  https://review.openstack.org/8268317:25
*** andreaf has joined #tripleo17:25
*** imtiaz has joined #tripleo17:32
*** jprovazn has quit IRC17:32
*** e0ne has quit IRC17:36
imtiazHi all. I'm wondering if anyone can give me some pointers on how to boot a baremetal server from a qcow2 image. From what I understand PXE boot process by default doesn't know how to deal with a qcow2 file and expects a kickstart file.17:36
Ngimtiaz: the qcow2 is never booted via PXE. Instead the machine is booted into a deployment image which exports the machine's disk to the server controlling things, so it can write the qcow2 image to the machine/s disk17:41
Ngmachine reboots, boots from disk, bam.17:41
*** mrunge has quit IRC17:41
Lotus907efiwhat is the possibility of adding a variable to the diskimager-builder line in devtest_overcloud.sh that builds the end user test image so it would allow setting the variable to an element list?17:45
*** hashar has quit IRC17:47
NgLotus907efi: doesn't seem beyond the realms of possibility, but really that image is just booted as a "look ma, it really works!" step17:47
Lotus907efiyes, but it builds by default from archive.ubuntu.blah17:47
*** epim has joined #tripleo17:48
Lotus907efiwhat I did last night was add the "apt-sources" element to that line17:48
*** mestery_ has joined #tripleo17:48
derekhLotus907efi: like this https://review.openstack.org/#/c/82739/217:48
*** mestery has quit IRC17:48
Lotus907efiso it would allow me to let the system build the image from my internal ubuntu repository17:48
Lotus907efiYes, that is what I meant.17:49
*** athomas has quit IRC17:49
Lotus907efiany idea why Dan Prince was against merging it???17:49
Lotus907efiit seems like a reasonable thing to me17:50
*** mestery_ is now known as mestery17:50
imtiaz+Ng: Thanks for your answer. So this deployment image has to be first created with disk-image-builder ?17:50
Ngimtiaz: correct. tripleo-incubator/scripts/devtest_ramdisk.sh is where we do it in the devtest.sh story17:50
*** tzumainn has quit IRC17:52
imtiaz+Ng: Excellent. Thanks again. I'll look it up.17:53
NgLotus907efi: so if that's all you want to do specifically, export NODE_DIST="ubuntu apt-sources" - that will affect all the image builds, including the demo image17:53
Lotus907efiyeah, lifeless told me that too17:54
*** mrunge has joined #tripleo17:55
*** derekh has quit IRC17:55
*** imtiaz has quit IRC17:57
*** pbelanyi has quit IRC18:00
*** eguz has joined #tripleo18:02
Lotus907efihmm, devtest run last night seems to have resulted in a not-compute VM doing a kernel oops when trying to bootup18:04
greghaynesKnown bug18:04
Lotus907efihmm, ok18:04
greghaynesshould be fairly rare18:04
Lotus907efiso delete the qcow image and rebuild the image, or do I need to pull a fresh code tree with fixes?18:05
*** eghobo has quit IRC18:05
*** Matt2 has left #tripleo18:06
greghaynesjust delete image and rebuild18:07
SpamapSsaurabhs1: yes that is perfect thank you18:09
SpamapSlsmola_: host software config will eventually help us not need to have explicit wait conditions18:09
*** pblaho has quit IRC18:09
SpamapSlsmola_: but it suffers the same problem with updates that regular wait conditions / handles suffer18:10
*** vkozhukalov has joined #tripleo18:10
*** Matt1 has joined #tripleo18:16
*** jang1 has joined #tripleo18:21
*** shakamunyi has joined #tripleo18:25
*** shakamunyi has joined #tripleo18:26
*** shakamunyi has quit IRC18:26
*** mrunge has quit IRC18:26
*** shakamunyi has joined #tripleo18:27
*** shakamunyi has quit IRC18:27
dprincelucasagomes: -2 says boom like nothing else. :) heh18:28
lucasagomesdprince, heh yeah sorry for that18:28
dprincelucasagomes: no worries man, glad we got it sorted out somewhere18:28
lucasagomesdprince, yeah indeed!18:29
lucasagomesbooooooom! :)18:29
*** panda has quit IRC18:29
openstackgerritStuart McLaren proposed a change to openstack/tripleo-image-elements: Glance: Create glance.protocol param for http/https  https://review.openstack.org/8289018:30
dprincelucasagomes: BTW, your going down for that. In foosball that is18:30
lucasagomesdprince, hah, well you better start practicing for atlanta then :P18:31
dprincelucasagomes: nah man, practice is for rookies. I'm just going to use my left hand this time (goelie shots)18:32
* dprince activates brick wall18:32
lucasagomesslagle, we need to play again against dprince and derekh, let's beat them this time18:33
dprincelucasagomes: oh, so on a more serious note. Do you have a minute to chat about the Ironic iboot driver patch?18:35
lucasagomesdprince, sure18:35
dprincelucasagomes: so what needs doing?18:35
*** ifarkas_ has quit IRC18:35
lucasagomesdprince, I got rebase that patch and fix the problems addressed by the reviewers18:35
jang1evening all18:37
dprincelucasagomes: okay, last rebase was a week ago?18:37
jang1lifeless: saw that pypi extra-index-url patch, it's made quite a difference.18:37
lifelessjang1: sweet18:38
lifelessjang1: science, who knew.18:38
lucasagomesdprince, no, I think december or something heh I totally abandoned that patch because I was working on other stuff18:38
lucasagomesdprince, so you need it soon? cause I don't think it will land before juno18:39
lucasagomesI can try to fix the problems and rebase it tomorrow18:40
dprincelucasagomes: I would like it in (because I use it) :) Don't care when but sooner is better than later although I can wait for Juno NP.18:40
lucasagomesdprince, ack18:41
*** e0ne has joined #tripleo18:41
lucasagomesI will at least fix the major problem pointed to the people on that review18:41
Lotus907efijang1: which patch is that for the pypi extra-index-url ? I would like to take a look at that too18:41
lifelessLotus907efi: ee18fe41da6c77b2ef0e1add4a6feb93b9a31ac8 in diskimage-builder18:42
jang1It's landed. A new tunable in the pypi element - DIB_NO_PIP_something.18:42
dprincelucasagomes: cool.18:42
dprincelucasagomes: I'll try and chip in some too because this is most important to me at this point anyways18:43
lifelessLotus907efi: just re-read the element/pypi/README.md18:43
dprincelucasagomes: appreciate the work you put into porting this one already18:43
lifelessLotus907efi: it also permits you to use wheels (assuming a suitable mirror is present)18:43
lucasagomesdprince, ack :) cheers for that18:43
lifelessjang1: I can now build *and deploy* an overcloud in 17m18:44
lifelessjang1: with --offline -u18:44
Lotus907efilifeless: mordred was setting up a pypi mirror here in Fort Collins a week or two ago. Do you know the URL for that?18:44
lifelessmordred: ^18:44
lifelessLotus907efi: I don't, but hopefully I've just invoked him :)18:45
Lotus907efimagic invocation of mordred?   cool trick that18:45
jang1Lotus907efi: somewhere on the wiki page there's a summary of site-local mirrors18:46
Lotus907efiyou ARE a wizard of higher order18:46
jang1I think the ftc one is filled in18:46
Lotus907efiummm, which wiki page? openstack or hpcloud?18:46
Lotus907efinuts, status of FtC mirror is marked: blocked waiting on more storage18:47
jang1have a quick whip-round the office :-)18:48
*** lucasagomes is now known as lucas-afk18:48
*** pbelanyi has joined #tripleo18:48
*** rpodolyaka1 has joined #tripleo18:50
*** shakayumi has joined #tripleo18:50
Lotus907efiso my pypi mirror is only using under 400MB so that seems like it should not be a huge problem18:51
*** shakayumi has joined #tripleo18:51
Lotus907efimordred: any news about the FtC pypi mirror system? Is it even semi-functional?18:52
*** ifarkas_ has joined #tripleo18:53
*** ramishra has quit IRC18:55
*** akrivoka has quit IRC18:55
*** ramishra has joined #tripleo18:55
*** derekh has joined #tripleo18:59
*** jistr has joined #tripleo18:59
*** ramishra has quit IRC18:59
lifelessmeeting time, right ?19:01
*** jprovazn has joined #tripleo19:02
*** shakayumi has quit IRC19:05
*** noslzzp has quit IRC19:06
*** hewbrocca has quit IRC19:09
*** rpodolyaka1 has quit IRC19:11
*** rpodolyaka1 has joined #tripleo19:11
*** slagle_ has joined #tripleo19:13
*** slagle has quit IRC19:15
*** noslzzp has joined #tripleo19:15
*** hewbrocca has joined #tripleo19:17
*** openstackgerrit has quit IRC19:18
*** openstackgerrit has joined #tripleo19:18
*** slagle_ is now known as slagle19:18
openstackgerritJon-Paul Sullivan proposed a change to openstack/tripleo-incubator: Add the select-cloud script  https://review.openstack.org/8280719:26
*** jp_at_hp has quit IRC19:27
mordredLotus907efi: sorry, it's only half functional. maybe cody-somerville will fix it19:28
*** pblaho has joined #tripleo19:29
*** pblaho has quit IRC19:31
*** hashar has joined #tripleo19:38
lifeless[2014-03-26 08:20:14] Total runtime: 2181 s19:43
lifeless[2014-03-26 08:20:14]   ramdisk    : 61 s19:43
lifeless[2014-03-26 08:20:14]   seed       : 412 s19:43
lifeless[2014-03-26 08:20:14]   undercloud : 545 s19:43
lifeless[2014-03-26 08:20:14]   overcloud  : 1144 s19:43
lifelesstotal devtest.sh runtime - 36m19:43
*** shakayumi has joined #tripleo19:45
*** shakayumi has quit IRC19:45
GheRiverolifeless: I want your machine19:46
Lotus907efimordred: I can probably fix it if no-one else is available19:48
lifelessGheRivero: thats wheels + --offline + -u19:48
greghayneswheels were huge for me19:49
greghaynesI should try the other two..19:49
lifelessGheRivero: plus an model name      : Intel(R) Core(TM) i7-4770K CPU @ 3.50GHz19:49
Lotus907efiyeah, even with local mirrors now functioning for me I am still just over an hour on a devtest run19:49
Lotus907efiand this is on a 24 core system with 24GB or memory19:49
Lotus907efiof memeory19:49
* Lotus907efi can't type19:50
GheRiverolifeless: disks?19:50
Lotus907efilifeless: I looked at the pypi README.md file and I do not see a mention of -u19:50
Lotus907efiwhat does that do?19:51
lifelessGheRivero: spinning platters19:51
lifelessGheRivero: just ext419:51
openstackgerritA change was merged to openstack/tripleo-image-elements: Configure uri_v3 auth url  https://review.openstack.org/8203519:51
GheRiveroI don't get it. Same CPU, 16GB RAM, SSD+HDD and overcloud times out19:52
*** shakayumi has joined #tripleo19:54
*** ramishra has joined #tripleo19:56
*** shakamunyi has joined #tripleo19:56
Lotus907efiGheRivero: what does  heat event-list overcloud tell you?19:56
Lotus907efiit might tell you why overcloud is timing out19:56
GheRiveroLotus907efi: retrying. Last error was cloud-init kernel pacnic trying to resize19:57
*** shakayumi has quit IRC19:58
greghaynesGheRivero: ah, thats the known random resize fail :/19:58
greghaynesalthough I couldnt find a ticket for it when I looked earler...19:59
GheRiverobut only with amd64 guests. With i386 it just timeout19:59
greghaynesOh, you consistently get resize fail with amd64?19:59
GheRiveroI would say yes20:00
greghaynesodd. I was getting that ~1/100 builds... maybe you have a nice setup to debug it :)20:00
GheRiverolifeless: are you building i386 or amd64 guests?20:00
lifelessGheRivero: amd6420:01
lifelessGheRivero: since thats what my host machine has and thus the wheels are20:01
lifelessI've given dprince a copy of a qcow2 file that shows the resize bug20:01
* tchaypo wakes up just in time to miss the meeting20:01
*** CaptTofu has joined #tripleo20:02
Lotus907efiresize bug only shows up for the overcloud control VM or the compute VMs too?20:02
dprincelifeless: when you built this image were you using the latest DIB, also: did you the naughty --max-online-resize option?20:03
*** shakamunyi has quit IRC20:03
Lotus907efiI just had a overcloud control node fail to boot because of that (I think it was the resize bug anyway)20:03
*** pbelanyi has quit IRC20:04
lifelessdprince: no20:04
mordredLotus907efi: awesome. as soon as I'm out of these meetings, I'll send you a message on where the stuff is20:04
lifelessdprince: thats why I pinged you, because we know now that that option isn't the only cause of the issue20:04
dprincelifeless: well, I'd like to see the full stack trace. But sure, maybe there is something else that exposes it.20:05
*** derekh has quit IRC20:05
*** lblanchard has quit IRC20:05
lifelessSpamapS: / slagle: https://review.openstack.org/#/c/82677/3 <- pre-req for Ironic20:05
dprincelifeless: like I said though the kernel side fix should be in route (via kernel updates)20:05
lifelessdprince: I'd love to see it too, haven't had time to slap a serial console on the vm20:06
*** jistr has quit IRC20:06
*** lblanchard has joined #tripleo20:06
Lotus907efiI got an error of "Warning: Failed to resize filesystem"  is that the same thing?20:06
*** shakamunyi has joined #tripleo20:06
*** rpodolyaka1 has quit IRC20:07
openstackgerritA change was merged to openstack/tripleo-image-elements: Permit overriding the nova compute manager.  https://review.openstack.org/8267720:09
*** rpodolyaka1 has joined #tripleo20:11
*** e0ne has quit IRC20:15
*** rpodolyaka1 has quit IRC20:15
*** jprovazn has quit IRC20:15
*** e0ne has joined #tripleo20:16
bnemecAnother easy +A for someone: https://review.openstack.org/#/c/80567/20:17
openstackgerritA change was merged to openstack/tripleo-image-elements: Add a softlink for /var/log/mysql  https://review.openstack.org/8056720:18
openstackgerritA change was merged to openstack/tripleo-image-elements: Custom service file is not needed for qpidd on systemd based hosts  https://review.openstack.org/8190120:19
*** e0ne has quit IRC20:20
openstackgerritA change was merged to openstack/diskimage-builder: DHCP: make udev rules want dhcp-interface@.service  https://review.openstack.org/8159620:23
*** rlandy has quit IRC20:24
*** vkozhukalov has quit IRC20:28
*** dprince has quit IRC20:33
openstackgerritBen Nemec proposed a change to openstack/tripleo-image-elements: Add no-op cinder-tgt element  https://review.openstack.org/8186720:33
openstackgerritBen Nemec proposed a change to openstack/tripleo-image-elements: Add mysql-mariadb-compat element  https://review.openstack.org/7803420:33
openstackgerritBen Nemec proposed a change to openstack/tripleo-image-elements: Add cinder-lio element  https://review.openstack.org/7846320:33
openstackgerritBen Nemec proposed a change to openstack/tripleo-image-elements: Factor out tgt-specific parts of cinder element  https://review.openstack.org/7846220:33
openstackgerritBen Nemec proposed a change to openstack/tripleo-image-elements: Allow os-svc-daemon to specify dir name with -d option  https://review.openstack.org/7803320:33
*** rpodolyaka1 has joined #tripleo20:38
openstackgerritJames Slagle proposed a change to openstack/tripleo-image-elements: Add fedora-rdo-icehouse element  https://review.openstack.org/8065820:39
openstackgerritJames Slagle proposed a change to openstack/tripleo-image-elements: Enable os-collect-config for the package install  https://review.openstack.org/8192720:40
openstackgerritA change was merged to openstack-infra/tripleo-ci: Remove "Revert https://review.openstack.org/#/c/58533/19"  https://review.openstack.org/8246220:42
*** rpodolyaka1 has quit IRC20:42
lifelessSpamapS: slagle: can I get a hell yeah on https://review.openstack.org/#/c/82414/ ?20:44
slaglelifeless: i'll trade you for one: https://review.openstack.org/#/c/81927/20:46
*** ramishra has quit IRC20:47
*** ramishra has joined #tripleo20:48
openstackgerritA change was merged to openstack/tripleo-heat-templates: Fix Ironic templates  https://review.openstack.org/8241420:48
openstackgerritA change was merged to openstack/tripleo-image-elements: Enable os-collect-config for the package install  https://review.openstack.org/8192720:52
*** ramishra has quit IRC20:52
openstackgerritA change was merged to openstack/diskimage-builder: Set +x on executable files  https://review.openstack.org/8163320:54
lifelessjogo: H201 <- WAT? thats so Python 2.3 man20:55
jogolifeless: you want to use 'except:'?20:57
lifelessjogo: I want to know the rationale for objecting to it21:01
*** jang1 has quit IRC21:01
lifelesssince in dead unsupported Pythons it was a problem but is not for anything recent (2.4?2.5? + )21:02
lifelessgreghaynes: https://review.openstack.org/#/c/81959/21:02
jogolifeless: go on ...21:02
lifelessjogo: so old bare except would catch internal-to-the-vm exceptions21:02
jogoalso there was a massive ML thread on this at one point21:02
lifelessjogo: like SystemExit, KeyboardInterrupt and GeneratorExit21:03
jogoahh so that part of the rule may be outdated21:03
lifelessjogo: this was fixed because it was a huge bug. Now it does the right thing.21:03
jogoin general though we don't want 'except' OR 'except Excetpion'21:03
jogoas both are too general21:03
clarkbpython 3 requires you not have a bare except right?21:04
clarkbI thought that was the rational21:04
lifelessclarkb: ahahaha no.21:04
lifeless>>> try:21:04
lifeless...    1/021:04
lifeless... except:21:04
lifeless...    pass21:04
lifelessPython 3.3.1 (default, Apr 17 2013, 22:30:32)21:04
lifelessjogo: except: is identical to except Exception:21:05
lifelessjogo: with one special case - it catches string exceptions.21:05
lifelessjogo: BUT21:05
lifelessjogo: they are gone in 3, so it really can't matter for 'us21:05
lifelessjogo: I don't get 'too general'21:05
lifelessjogo: since try:finally: is 'do something after errors but raise'21:06
lifelessjogo: try:except: is 'do something after errors but I may choose not to raise'21:06
jogolifeless: in python 2.7.5 a KeyboardInterrupt is caught by 'except:' and not 'except Exception'21:06
* lifeless cross checks21:07
jogotry: time.sleep(10)21:07
jogoexcept Exception: print "OH HAI"21:07
lifelessyou're right, I'm clearly wrong.21:08
lifelessjogo: ok, ignore my whinge.21:08
*** shakayumi has joined #tripleo21:08
greghayneslifeless: responded21:08
jogoNever use catch-all except: statements, or catch Exception or StandardError, unless you are re-raising the exception or in the outermost block in your thread (and printing an error message). Python is very tolerant in this regard and except: will really catch everything including misspelled names, sys.exit() calls, Ctrl+C interrupts, unittest failures and all kinds of other exceptions that you simply don't want to catch.21:09
*** shakayumi has quit IRC21:09
lifelessgreghaynes: oh, I fixed that locally, but failed.21:09
* jogo goes back to ignoring lifeless21:09
Lotus907efilifeless: yesterday you said there was a simple option change to apt-mirror to create a single unified tree instead of the separate tree that jeb did21:10
Lotus907efiI want to try setting up a precise mirror so I can try precise overcloud images21:10
*** shakamunyi has quit IRC21:11
openstackgerritlifeless proposed a change to openstack/tripleo-incubator: Add Ironic support to incubator.  https://review.openstack.org/8195921:11
openstackgerritlifeless proposed a change to openstack/tripleo-incubator: Remove backwards compatability in setup-baremetal  https://review.openstack.org/7846721:11
lifelessLotus907efi: http://paste.ubuntu.com/7153336/21:12
Lotus907efiok, thanks21:12
lifelessLotus907efi: thats my mirror of saucy amd64 only21:12
lifelesserm I mean amd64 + i386 but not all components21:12
Lotus907efiok, I can alter for precise21:13
lifelessin that I skip backports and proposed21:13
Lotus907efiand that cust down how much you have to mirror?21:13
lifelesssome, but its just stuff I don't want21:13
lifelessI don't use backports or proposed often enough to want it21:13
lifelessyou could skip source if you want to reduce the mirror load a lot21:14
lifelessfor instance21:14
*** jp_at_hp has joined #tripleo21:14
*** jp_at_hp has joined #tripleo21:14
*** ifarkas has quit IRC21:15
lifelessslagle: one more - https://review.openstack.org/#/c/82431/21:15
*** ifarkas_ has quit IRC21:16
lifelessslagle: this will let us change toci to make the calls needed to let us move forward with the refactor needed for ironic....21:16
*** noslzzp has quit IRC21:17
jp_at_hpSpamapS, Ng tests have passed on https://review.openstack.org/#/c/80644, any chance you could re-review please?21:17
*** jp_at_hp has quit IRC21:19
*** lblanchard has quit IRC21:21
lifelessLotus907efi: https://bugs.launchpad.net/ubuntu/+source/linux/+bug/1233075 - what DIB_RELEASE did you use ?21:28
uvirtbotLaunchpad bug 1233075 in linux "resizefs failure with raring" [High,New]21:28
Lotus907efiI did not set a DIB_RELEASE at all21:28
Lotus907efiso it would have used saucy by default21:28
Lotus907efiso this is not really with raring at all21:28
lifelessLotus907efi: getting a serial console trace to see the top of the stacktrace would be valuable21:29
Lotus907efiok, how do I do that with a VM? use virsh?21:30
lifelessuse virt-manager to add a serial console device, then virsh or virt-manager should let you connect to it, then power off/power on the vm21:30
Lotus907efioh, nuts I already did "heat stack-delete overcloud" in preparation for building a precise overcloud21:31
Lotus907efiI guess I could just re-run devtest_overcloud.sh with no changes21:31
Lotus907efiand see if it does the same thing again21:31
bnemecAs long as you don't rebuild the image I think it will fail the same way.21:32
bnemecAt least it always has for me.21:32
*** akrivoka has joined #tripleo21:33
*** hashar has quit IRC21:33
Lotus907efiok, re-running devtest_overcloud.sh with no changes21:34
lifelessLotus907efi: use USE_CACHE=121:41
lifelesse.g. export USE_CACHE=121:41
lifelessor whatever21:41
lifelesswill keep the image you have and should reproduce the error21:41
*** UtahDave has joined #tripleo21:41
Lotus907efitoo late21:42
*** UtahDave has left #tripleo21:42
*** matty_du1s has joined #tripleo21:42
*** jeblair_ has joined #tripleo21:42
*** jeblair_ is now known as corvus21:43
*** jeblair has quit IRC21:45
*** matty_dubs has quit IRC21:45
*** corvus is now known as jeblair21:45
*** notq has quit IRC21:46
*** StevenK has quit IRC21:47
*** al has quit IRC21:47
*** zigo has quit IRC21:47
*** wendar has quit IRC21:47
*** StevenK has joined #tripleo21:47
*** al has joined #tripleo21:47
*** notq has joined #tripleo21:49
*** wendar has joined #tripleo21:49
*** zigo has joined #tripleo21:49
lifelessStevenK: morning21:49
lifelessStevenK: do you have the toci patch up to use --nodes ?21:49
lifelessSpamapS: StevenK: lsmola_: https://review.openstack.org/#/c/82431/21:50
lifelessslagle: https://review.openstack.org/#/c/82431/21:50
lifelessNg: if you're still conscious ^21:50
Lotus907efimordred: are you still in meetings? I am about to run out and get some lunch but will be back shortly21:56
*** markmc has quit IRC22:01
lifelessjogo: OH, I remember. Exception used to include everything. Then BaseException was added and it all got better.22:02
* lifeless is clearly in his dotage.22:02
*** hewbrocca has quit IRC22:03
*** sseago has quit IRC22:03
StevenKlifeless: Not yet, I was planning to write that after testing my setup-baremetal change22:09
lifelesssoon as slagle or SpamapS get back I hope to get https://review.openstack.org/#/c/82431/ landed22:25
*** matty_du1s is now known as matty_dubs|gone22:26
tchaypoalmost morning22:28
lifelesstchaypo: what time is it?22:29
tchaypoalmost coffee time. i got dragged out for breakfast "at 9" and we're only just sitting down now22:30
lifelessso its 930 ?22:31
tchaypoyeah, and i'm reading my email while i wait for coffee and food22:34
lifelessoh sadface, xtranormal is gone22:35
tchaypooh no22:36
tchaypoone of my workmates ago G used it to great effect22:36
tchaypoactually many of them did, but one in particular22:36
lifelessGheRivero: how did your virtio test go ?22:40
StevenKlifeless: I was trying to test the new style call for setup-baremetal, but my seed does not love me22:40
lifelessStevenK: No?22:42
StevenKlifeless: You were helping yesterday22:42
lifelessStevenK: I am confuse22:43
GheRiverolifeless: Still figthing with the cloud-init resize kernel panics, but doesn't look to improve things to much22:43
* tchaypo reads meeting minutes22:43
lifelessGheRivero: :(22:44
lifelessGheRivero: resize issues - try tweaking the 0.X ratio a bit, it seems to let one avoid tickling it22:44
lifelessGheRivero: or perhaps try using --max-resize22:45
StevenKlifeless: I was trying to test the new way to run setup-baremetal yesterday before working on the toci patch, but my seed isn't listening on, and I couldn't work out how to fix it.22:45
tchaypoI don't think triage is hard/annoying, it's just easy to one of those important-but-not-urgent small tasks that are easy to forget when you have urgent things in your face22:45
lifelessStevenK: ah, ok. so.22:45
StevenKlifeless: Anyway, I have a toci patch locally22:45
GheRiverowill do when the current build finish.22:45
lifelessStevenK: log into the seed. Look at os-collect-config.22:45
GheRiverojust need to wait.... 30 minutes :/22:46
lifelessGheRivero: you're doing a full run?22:46
GheRiveronow yes22:48
StevenKlifeless: The logs of os-collect-config are from Monday, which is odd22:48
GheRiveroanyhow, I guess I'm hitting some disk problem in the lab server. Almost two minutes just to build the ramdisk doesn't look right22:49
lifelessGheRivero: thats pretty slow22:50
lifeless60s for me22:50
Lotus907efinuts, overcloud control node failed to create but in a different way this time22:52
*** jpeeler has quit IRC22:54
lifelessLotus907efi: how did it fail ?22:54
lifelessStevenK: try os-collect-config --force --one22:54
Lotus907efinot finding a network to pxe boot on when I try to boot the image22:55
* StevenK tries to work what his local changes to the no-more-backward-compat branch is so he can reconcile22:55
StevenKlifeless: Ah ha, that got it working22:56
lifelessSpamapS: what do you think of a reboot triggering os-c-c even if no changes found ?22:56
lifelessSpamapS: in the cache22:56
openstackgerritBen Nemec proposed a change to openstack/diskimage-builder: POC: Add ability to run test scripts  https://review.openstack.org/8293922:56
openstackgerritBen Nemec proposed a change to openstack/diskimage-builder: Add tests.d to cache-url  https://review.openstack.org/8294022:56
SpamapSlifeless: +122:57
StevenKAh ha, I'm trying to address review comments, and lifeless is doing something else22:57
SpamapSlifeless: I think there's a difference between "I have different metadata in my cache" and "I have not yet applied my metadata to the system."22:58
SpamapSlifeless: In theory things should set up boot-time configurations outside of orc scripts.22:58
SpamapSlifeless: but in practice, we're obviously doing it wrong.22:58
StevenKAh, it's a rebase. I think.22:58
*** epim has quit IRC23:01
openstackgerritSteve Kowalik proposed a change to openstack-infra/tripleo-ci: Use --nodes for setup-baremetal in toci_devtest  https://review.openstack.org/8294123:01
lifelessSpamapS: well, not sure I entirely agree with that 'in theory'23:03
lifelessSpamapS: I like the idea of starting from 'we booted and got network' each time, for consistency23:04
lifelessSpamapS: because new images will always wipe out a lot of stuff23:04
lifelessStevenK: yes I rebased to fix review issues in the patch above - sorry, I hadn't gotten to the ones below23:05
openstackgerritA change was merged to openstack/tripleo-incubator: Support for specifying nodes for setup-baremetal  https://review.openstack.org/8243123:05
lifeless\o/ thanks GheRivero !23:07
lifelessStevenK: can you re-push https://review.openstack.org/#/c/82941/ with a trivial change - for zuul to start the check jobs over?23:07
SpamapSlifeless: perhaps we should start pushing harder for readonly root23:08
SpamapSlifeless: that would prevent the false sense of survivability :-P23:08
SpamapSlifeless: also if we do a "first boot gets automatic force" .. we may want to be more careful about when we run it23:08
StevenKlifeless: Or recheck no bug ?23:09
SpamapSok so three times I've failed to boot something on cd-undercloud23:10
SpamapSwe may have used up all the boxes :-/23:10
SpamapSmight need to start aggressively pursuing replacements/fixes/etc.23:11
openstackgerritSteve Kowalik proposed a change to openstack-infra/tripleo-ci: Use --nodes for setup-baremetal in toci_devtest  https://review.openstack.org/8294123:11
greghaynesIs there a trick for quick feedback loop of messing with a heat template -> testing what the metadata api json will end up being23:11
lifelessStevenK: recheck no bug doesn't cancel current checks *AFAIK*23:12
greghaynesbasically resolve all references, fn::*, etc23:12
lifelessgreghaynes: heat stack-update ?23:13
lifelessgreghaynes: should be 3-4 seconds, then os-collect-config --print23:13
greghaynesso obvious in hindsight23:13
Lotus907efianyone ever see "Buffer I/O error on device loop0p1, logical block 0" or "Buffer I/O error on device loop0p1, logical block 401623" on the real underlying HW system that is running devtest.sh ?23:13
lifelessgreghaynes: I even wrote it in the draft docs23:14
lifelessLotus907efi: its udev poking at the raw device23:14
lifelessLotus907efi: after its gone23:14
lifelessLotus907efi: AFAIK23:14
Lotus907efiit's normal?23:14
lifeless*AFAIK* its an armless race condition23:14
SpamapSLotus907efi: yes23:16
*** andreaf has quit IRC23:22
*** cwolferh has quit IRC23:27
*** giulivo has quit IRC23:28
*** julim has quit IRC23:30
*** cwolferh has joined #tripleo23:34
openstackgerritDerek Higgins proposed a change to openstack/tripleo-incubator: Document remote-host in test environment json  https://review.openstack.org/8294623:43
*** geerdest has quit IRC23:46
*** noslzzp has joined #tripleo23:51
*** lucas-afk has quit IRC23:52
tchaypoI'm really getting sick of "git.openstack.org: Temporary failure in name resolution"23:55
tchaypoI'm starting to think we can open a bug for that one just so that we can get better tracking of how often it bites23:57
lifelessStevenK: does https://jenkins04.openstack.org/job/check-tripleo-seed-precise/285/ look like its running your --nodes code?23:57
lifelessStevenK: if so, we can land I think23:57
lifelesstchaypo: tell -infra, they have been trying to reduce dns issues23:57
StevenKThe seed doesn't run setup-baremetal23:58
lifelessStevenK: sure about that ?23:58
lifelessoh, from toci. righto.23:59
StevenKIt runs setup-baremetal, but that's incubator code, not the call we're both interested in23:59

Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!