Wednesday, 2013-12-04

*** rongze has quit IRC00:04
SpamapS30878 root      20   0 19500 1884 1144 D     5  0.0   0:08.80 mkfs.ext3 -F -L ephemeral0 /dev/disk/by-path/ip-10.10.100:09
SpamapSinteresting.00:09
SpamapSnova-api is returning 500's for meta-data requests00:10
SpamapSbut no idea why00:10
SpamapS2013-12-04 00:11:36,393.393 993 ERROR nova.api.metadata.handler [req-a4bd8ff8-793e-4f15-93be-1c7572af45e9 None None] Failed to get metadata for ip: 10.10.16.17100:12
SpamapS2013-12-04 00:11:36,393.393 993 TRACE nova.api.metadata.handler   File "/opt/stack/venvs/nova/local/lib/python2.7/site-packages/nova/network/neutronv2/__init__.py", line 69, in get_client00:12
SpamapS2013-12-04 00:11:36,393.393 993 TRACE nova.api.metadata.handler     raise exceptions.Unauthorized()00:12
SpamapSok enough debugging for today00:12
* SpamapS -> kids00:12
*** boris-42 has quit IRC00:14
lifelessSpamapS: there was a fix for that00:16
lifelessSpamapS: I thought I pulled it in, may need to refresh nova again.00:16
lifelessSpamapS: which nova-api - the undercloud one?00:16
SpamapSlifeless: right, undercloud nova-api00:24
lifelessso yeah that was fixed00:27
lifelesslets see00:27
*** vipul is now known as vipul-away00:28
lifeless652620d12f3afe6845e41d9762b52d23f44fd55700:28
lifelessbut its not in /opt/stack/nova now00:28
SlickNikhello good folks.00:29
lifelessSlickNik: allo allo00:30
SlickNikI'm looking at moving the trove dib-image elementes into tripleo-image-elements, and I had two questions.00:30
*** cd-undercloud has joined #tripleo00:31
cd-undercloud************** overcloud complete status=1 ************00:31
*** cd-undercloud has quit IRC00:31
*** matsuhashi has joined #tripleo00:31
SlickNik1. We're still using first-boot.d. I know it's deprecated, and I will fix it over the next few days, but I'm wondering whether it's okay to do the move before fixing it.00:31
SlickNik2. Is there some way of moving the image-elements over while keeping the git history? I'm afraid if I do a regular gerrit submission, it will all be lost.00:32
lifelessSure, 1 is fine00:32
lifelessGit history will be lost.00:33
SlickNikFair enough. 2 is somewhat unfortunate, but understandable.00:33
SlickNikThanks lifeless!00:33
lifelessSpamapS: updated nova; did an os-collect-config --force --one; restarted tripleo-cd00:38
*** vipul-away is now known as vipul00:40
*** lucas-dinner has quit IRC00:45
*** openstack has joined #tripleo00:46
*** openstackgerrit has quit IRC00:56
*** openstackgerrit has joined #tripleo00:56
*** krotscheck has quit IRC01:19
*** julim has quit IRC01:22
*** noslzzp has joined #tripleo01:32
*** kui has joined #tripleo01:33
*** nosnos has joined #tripleo01:35
*** nosnos_ has joined #tripleo01:37
*** nosnos has quit IRC01:40
*** cd-undercloud has joined #tripleo01:40
cd-undercloud************** overcloud complete status=1 ************01:40
*** cd-undercloud has quit IRC01:40
*** epim has quit IRC01:44
openstackgerritA change was merged to openstack/tripleo-incubator: Add Ironic to setup-undercloud-passwords  https://review.openstack.org/5980002:00
dkehnlifeless: are you going to around for a bit02:01
*** rongze has joined #tripleo02:01
*** rongze has joined #tripleo02:02
lifelessdkehn: ues02:13
dkehnlifeless: I'm in the middle of kids home work, but just wondering if your going to be around later, I'm seeing an interesting situation were the MAC is being chopped 1 octect short in register-nodes02:14
lifelessdkehn: check pending reviews - there is a review there to fix a bug in a wait_for loop.02:15
dkehnyepper that would be it02:15
dkehnyepper hard coded02:16
*** rwsu has quit IRC02:35
*** rwsu has joined #tripleo02:39
*** cd-undercloud has joined #tripleo02:42
cd-undercloud************** overcloud complete status=1 ************02:42
*** cd-undercloud has quit IRC02:42
SpamapSooo02:45
SpamapSthe stack actually did complete02:46
SpamapSjust took a really long time02:46
*** marun has joined #tripleo02:46
SpamapS| CompletionHandle    | 80939 | state changed          | CREATE_IN_PROGRESS | 2013-12-04T02:04:04Z |02:46
SpamapS| CompletionCondition | 80962 | state changed          | CREATE_COMPLETE    | 2013-12-04T02:43:19Z |02:46
SpamapS| notcompute          | 80953 | state changed          | CREATE_IN_PROGRESS | 2013-12-04T02:04:09Z |02:55
SpamapS| notcompute          | 80956 | state changed          | CREATE_COMPLETE    | 2013-12-04T02:30:44Z |02:55
SpamapS26 minutes to do two boxes02:55
*** ruhe has joined #tripleo02:56
lifelessok, so what was the failure? timeout?02:57
SpamapSlifeless: yes02:58
*** ruhe has quit IRC03:01
*** vkozhukalov has joined #tripleo03:09
SpamapSlifeless: it is going now .. might be good to monitor and see how the dd performs.. anything else..03:09
lifelessinstrumentation; we did a design session o nit03:10
*** vipul has quit IRC03:15
*** shadower_ has joined #tripleo03:16
*** vipul has joined #tripleo03:17
*** shadower has quit IRC03:17
*** rainya has quit IRC03:17
*** antonym has quit IRC03:18
*** rainya has joined #tripleo03:19
*** phschwartz has quit IRC03:19
*** antonym has joined #tripleo03:21
*** phschwartz has joined #tripleo03:22
*** cd-undercloud has joined #tripleo03:46
cd-undercloud************** overcloud complete status=0 ************03:46
*** cd-undercloud has quit IRC03:46
SpamapS\o/03:52
SpamapSjust under the wire though03:53
*** morazi has quit IRC03:53
*** rbrady has quit IRC04:04
*** CaptTofu has quit IRC04:15
*** CaptTofu has joined #tripleo04:16
SpamapSoohh04:24
SpamapSI didn't realize we had saucy by default now04:24
SpamapScool!04:24
SpamapSbut04:24
SpamapShave to wonder if something regressed performance wise.04:24
*** SpamapS changes topic to "Using OpenStack to deploy OpenStack; meetings Tuesday 1900 UTC in #openstack-meeting-alt"04:25
lifelessSpamapS: well done 86!04:30
*** rbrady has joined #tripleo04:34
*** nosnos has joined #tripleo04:38
*** nosnos_ has quit IRC04:42
*** rongze has quit IRC04:45
*** StevenK has joined #tripleo04:47
*** cd-undercloud has joined #tripleo04:49
cd-undercloud************** overcloud complete status=1 ************04:49
*** cd-undercloud has quit IRC04:49
SpamapSnooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooo04:49
SpamapStimed out04:50
*** noslzzp has quit IRC04:51
openstackgerritClint "SpamapS" Byrum proposed a change to openstack/tripleo-incubator: Extend devtest_overcloud timeout by 5 minutes  https://review.openstack.org/5989504:53
*** rushiagr has joined #tripleo04:56
*** rbrady has quit IRC05:15
*** rongze has joined #tripleo05:16
*** rainya has quit IRC05:20
*** LinuxJedi_ has joined #tripleo05:20
*** michchap_ has joined #tripleo05:21
*** marun has quit IRC05:22
*** jayg|g0n3 has quit IRC05:22
*** stevebaker has quit IRC05:22
*** cwolferh has quit IRC05:22
*** michchap has quit IRC05:22
*** LinuxJedi has quit IRC05:22
*** phschwartz has quit IRC05:22
*** phschwartz has joined #tripleo05:22
*** antonym has quit IRC05:22
*** stevebaker has joined #tripleo05:22
*** marun has joined #tripleo05:22
*** cwolferh has joined #tripleo05:22
*** jayg|g0n3 has joined #tripleo05:23
*** rongze has quit IRC05:28
*** cd-undercloud has joined #tripleo05:51
cd-undercloud************** overcloud complete status=1 ************05:51
*** cd-undercloud has quit IRC05:51
lifelessSpamapS: well, we had our time in the sun05:56
SpamapSlifeless: 20:53 < openstackgerrit> Clint "SpamapS" Byrum proposed a change to openstack/tripleo-incubator: Extend  devtest_overcloud timeout by 5 minutes  https://review.openstack.org/5989505:56
SpamapSlifeless: we're _just barely_ missing it05:56
lifelessyah05:56
SpamapSlifeless: It is somewhat confusing why.. that instrumenting work would help.. but for now, I think we are just going to have to live with "it got slower"05:57
*** rongze has joined #tripleo05:58
*** rpodolyaka1 has joined #tripleo05:59
rpodolyaka1morning tripleo06:06
lifelessrpodolyaka1: o/06:21
*** kui has quit IRC06:21
*** rongze has quit IRC06:30
*** rongze has joined #tripleo06:30
*** nosnos has quit IRC06:34
*** nosnos has joined #tripleo06:35
*** jcooley_ has joined #tripleo06:41
*** akuznetsov has quit IRC06:44
*** jcooley_ has quit IRC06:51
*** boris-42 has joined #tripleo06:51
*** vkozhukalov has quit IRC06:51
*** cd-undercloud has joined #tripleo06:52
cd-undercloud************** overcloud complete status=0 ************06:52
*** cd-undercloud has quit IRC06:52
*** marun has quit IRC06:57
*** akuznetsov has joined #tripleo06:58
*** nosnos_ has joined #tripleo07:05
*** nosnos has quit IRC07:08
*** akuznetsov has quit IRC07:08
*** rpodolyaka1 has quit IRC07:16
*** antonym has joined #tripleo07:17
*** rwsu has quit IRC07:24
*** lsmola has joined #tripleo07:26
lsmolalifeless, hello, are you around?07:27
*** jtomasek has joined #tripleo07:29
*** jtomasek has quit IRC07:34
*** pblaho has joined #tripleo07:37
*** jtomasek has joined #tripleo07:37
*** rwsu has joined #tripleo07:48
*** cd-undercloud has joined #tripleo07:54
cd-undercloud************** overcloud complete status=1 ************07:54
*** cd-undercloud has quit IRC07:54
*** jprovazn has joined #tripleo07:55
*** bauzas has joined #tripleo08:08
*** marun has joined #tripleo08:09
*** jistr has joined #tripleo08:18
*** nosnos_ has quit IRC08:19
*** nosnos has joined #tripleo08:19
*** GheRivero has quit IRC08:22
*** rdopieralski has joined #tripleo08:26
*** martyntaylor has joined #tripleo08:28
*** vkozhukalov has joined #tripleo08:30
*** matsuhashi has quit IRC08:35
*** matsuhashi has joined #tripleo08:36
*** boris-42 has quit IRC08:40
*** martyntaylor has quit IRC08:46
*** cd-undercloud has joined #tripleo08:58
cd-undercloud************** overcloud complete status=1 ************08:58
*** cd-undercloud has quit IRC08:58
*** jcoufal has joined #tripleo08:58
*** martyntaylor has joined #tripleo09:00
Ngmorning09:00
*** athomas has joined #tripleo09:02
marios_o/09:02
openstackgerritA change was merged to openstack/tripleo-incubator: Extend devtest_overcloud timeout by 5 minutes  https://review.openstack.org/5989509:04
*** victor_lowther_ has joined #tripleo09:06
*** SlickN1k has joined #tripleo09:11
*** rpodolyaka has quit IRC09:12
*** funzo has quit IRC09:12
*** victor_lowther has quit IRC09:12
*** greghaynes has quit IRC09:12
*** zaro has quit IRC09:12
*** clarkb has quit IRC09:12
*** SlickNik has quit IRC09:12
*** SlickN1k is now known as SlickNik09:12
openstackgerritA change was merged to openstack/tripleo-incubator: Test MAC address should be 6 octets.  https://review.openstack.org/5981209:12
*** victor_lowther_ is now known as victor_lowther09:12
*** akuznetsov has joined #tripleo09:14
*** derekh has joined #tripleo09:15
*** zaro has joined #tripleo09:18
*** greghaynes has joined #tripleo09:18
*** clarkb has joined #tripleo09:19
*** rpodolyaka has joined #tripleo09:20
*** max_lobur_afk is now known as max_lobur09:24
*** lucasagomes has joined #tripleo09:31
*** markmc has joined #tripleo09:42
*** jtomasek has quit IRC09:43
*** jtomasek has joined #tripleo09:43
*** boris-42 has joined #tripleo09:46
*** akrivoka has joined #tripleo09:47
openstackgerritDerek Higgins proposed a change to openstack-infra/tripleo-ci: Add element to install a testenv worker  https://review.openstack.org/5830509:48
openstackgerritDerek Higgins proposed a change to openstack-infra/tripleo-ci: Add script to install a testenv client  https://review.openstack.org/5830609:48
*** cd-undercloud has joined #tripleo09:59
cd-undercloud************** overcloud complete status=0 ************09:59
*** cd-undercloud has quit IRC09:59
Ngooh, status=010:02
rdopieralski\o/10:02
*** nosnos_ has joined #tripleo10:03
openstackgerritJun Jie Nan proposed a change to openstack/diskimage-builder: Fix no busybox symlinks issue on rhel  https://review.openstack.org/5994010:06
*** nosnos has quit IRC10:06
*** matsuhashi has quit IRC10:22
*** athomas has quit IRC10:37
*** funzo has joined #tripleo10:43
*** athomas has joined #tripleo10:44
*** matsuhashi has joined #tripleo10:45
*** jistr has quit IRC10:46
*** rongze has quit IRC10:53
*** cd-undercloud has joined #tripleo11:02
cd-undercloud************** overcloud complete status=0 ************11:02
*** cd-undercloud has quit IRC11:02
*** jistr has joined #tripleo11:09
*** rongze has joined #tripleo11:11
*** jergerber has quit IRC11:43
*** rongze has quit IRC11:45
*** rongze has joined #tripleo11:51
*** max_lobur is now known as max_lobur_afk11:57
*** cd-undercloud has joined #tripleo12:05
cd-undercloud************** overcloud complete status=0 ************12:05
*** cd-undercloud has quit IRC12:05
*** CaptTofu has quit IRC12:30
*** CaptTofu has joined #tripleo12:31
*** jprovazn has quit IRC12:37
*** rbrady has joined #tripleo12:39
*** shadower_ is now known as shadower12:43
*** flwang has joined #tripleo12:43
jog0lifeless: turns out the gate is hurting again :/12:48
jog023% failure12:48
flwangmorning guys12:49
flwangmay I ask a stupid question about TripleO?12:49
*** max_lobur_afk is now known as max_lobur12:52
*** jcoufal has quit IRC12:52
*** akuznetsov has quit IRC12:53
*** akuznetsov has joined #tripleo12:56
*** jcoufal has joined #tripleo12:57
*** morazi has joined #tripleo12:57
*** jprovazn has joined #tripleo12:58
*** lucasagomes is now known as lucas-hungry12:59
rpodolyakaflwang: sure you may :)13:02
*** matsuhashi has quit IRC13:03
*** dprince has joined #tripleo13:03
*** panda_ has joined #tripleo13:03
*** panda_ is now known as panda13:04
*** cd-undercloud has joined #tripleo13:06
cd-undercloud************** overcloud complete status=42 ************13:06
*** cd-undercloud has quit IRC13:06
*** nosnos_ has quit IRC13:07
*** nosnos has joined #tripleo13:08
*** akuznetsov has quit IRC13:13
*** jdob has joined #tripleo13:14
*** CaptTofu has quit IRC13:15
*** CaptTofu has joined #tripleo13:16
*** akuznetsov has joined #tripleo13:16
*** rushiagr has quit IRC13:31
*** matsuhashi has joined #tripleo13:32
flwangis it possible to leverage TripleO to do some auto scaling?  as we said on the wiki https://wiki.openstack.org/wiki/TripleO  the cloud can be 'scaling down to as few as 2 machines'. thanks13:38
*** noslzzp has joined #tripleo13:38
*** noslzzp has quit IRC13:39
*** noslzzp has joined #tripleo13:39
*** nosnos has quit IRC13:39
*** CaptTofu has quit IRC13:39
jdobis there a way to see open code reviews for all of the tripleo-related projects, or do I need to do queries on tripleo-incubator, tripleo-image-elements, etc. individually?14:02
markmcjdob, see https://wiki.openstack.org/wiki/TripleO#Review_team14:03
*** lucas-hungry is now known as lucasagomes14:03
jdobi'm not familiar enough with the search syntax on the review site to know if it supports any so-- cool, will check there, thanks markmc14:03
markmcnp :)14:03
jdobah perfect, that's exactly what I was trying to do; I was missing the () around my attempt at an OR query14:03
*** julim has joined #tripleo14:05
rpodolyakajog0: hey, this might be interesting for you https://bugs.launchpad.net/openstack-ci/+bug/1216851/comments/614:06
uvirtbotLaunchpad bug 1216851 in nova "nova unit tests occasionally fail migration tests for mysql and postgres" [Undecided,Confirmed]14:06
jog0rpodolyaka: awesome sauce -- if we can't bump that number up for specific  tests lets bump it up accross the board14:07
jog0rpodolyaka: can you look at a working py27 test and see how long migrations usually take. are we alyways close to the timelimit or do we have some freak outliers14:08
rdopieralski>14:08
rdopieralski> Thanks,14:08
rdopieralski>14:08
rdopieralski> Sescia14:08
rdopieralskisorry14:08
rdopieralskimisclick14:08
*** cd-undercloud has joined #tripleo14:09
cd-undercloud************** overcloud complete status=0 ************14:09
*** cd-undercloud has quit IRC14:09
rpodolyakajog0: running test_walk_versions() on my machine, will tell you the results soon14:12
*** jayg|g0n3 is now known as jayg14:13
rpodolyakaflwang: sorry, missed you message. I would say, it will be possible to autoscale both overcloud and undercloud, but not right now14:14
*** rongze has quit IRC14:14
*** morazi has quit IRC14:15
jog0cool, you can also look at the logs from a working gate job14:16
*** rongze has joined #tripleo14:18
*** morazi has joined #tripleo14:26
rpodolyakajog0: so these are results for my machine http://paste.openstack.org/show/54424/14:30
rpodolyakajog0: checking gate logs14:31
rpodolyakajog0: interesting, I took 3 random successful runs of gate jobs in nova and results here vary greatly - e.g. test_mysql_opportunistically from 41s to 67s, test_mysql_opportunistically - 93s to 106s14:38
rpodolyakajog0: the latter was supposed to be test_postgresql_opportunistically14:38
rpodolyakajog0: oh, so for test_postgresql_opportunistically the best result I found was 38s.  So the results seem to depend heavily on the load of hosts running CI nodes (and postresql are often fairly close to 160s timeout we have now)14:42
*** matty_dubs|gone is now known as matty_dubs14:44
*** bauzas has quit IRC14:56
jog0rpodolyaka: cool it sounds like we should bump that timeout up then14:57
rpodolyakajog0: yeah, I wonder how much we should bump it though14:57
*** jtomasek has quit IRC14:57
*** fungi has joined #tripleo14:58
*** ChanServ sets mode: +v fungi14:58
fungijog0: sez someone might know why nova db migrations suddenly got slooooow14:58
jog0rpodolyaka: ^ fungi14:58
fungijog0: you mentioned a 160-second timeout. is that per migration step, or for the list of migrations in aggregate?14:59
rpodolyakafungi: all migrations14:59
fungioh, the 160sec timeout is a testr unit timeout, and the migrations are performed together as one test15:00
rpodolyakafungi: right15:00
fungiso devs are constantly adding new migration steps, and the timeout is now coming into sight on some runs as a result15:00
fungii assume the recommended solution for that is to perform each migration step as a separate test, with a known start and end state?15:01
*** jtomasek has joined #tripleo15:02
fungithat will presumably improve parallelization/throughput for the tests anyway15:02
*** jistr has quit IRC15:02
fungiassuming the migrations are able to be properly partitioned so that they use dedicated per-migration databases and don't step on one anothers' toes15:03
jog0fungi: good idea, rpodolyaka is that possible?15:03
*** jistr has joined #tripleo15:04
*** jistr is now known as jistr|mtg15:04
* rpodolyaka thinking15:04
mordredyou'd have to get the start state for the migration test15:04
mordredthe only currently known way to get a start state for a given migration is to run the migrations that come befor eit15:05
rpodolyakayeah15:05
mordredusually at the beginning of the cycle, someone replaces the migrations with a single rollup migration15:05
rpodolyakaso this seems to be a bad idea15:05
*** cd-undercloud has joined #tripleo15:05
cd-undercloud************** overcloud complete status=0 ************15:05
*** cd-undercloud has quit IRC15:05
rpodolyakaand we already have problems with tests accessing the same db which are run in parallel15:05
mordredwe shoudl really have schema-per-test15:06
rpodolyaka+115:06
* mordred needs to write the nova patch to use drizzle for unitttests instead of mysql15:06
*** jtomasek has quit IRC15:07
rpodolyakamordred: fungi: jog0: https://etherpad.openstack.org/p/db-schemas-provisioning-on-ci15:07
rpodolyakamordred: fungi: jog0: so we have a POC implemented in Nova/oslo-incubator, but clarkb said mysql openstack_citest user didn't have enough rights to create/drop schemas15:09
*** john-n-seattle1 has quit IRC15:09
mordredthat's right.15:11
fungirpodolyaka: i believe that is true for mysql but not for postgres15:11
mordredyou can't give create/drop schema in mysql without also giving admin rights, iirc15:12
mordredwhich is why  I'm mainly suggesting replacing mysql with drizzle in our testing, since you can just stsart drizzle with no auth plugins and there will be no auth15:12
*** jtomasek has joined #tripleo15:12
* rpodolyaka googling15:12
mordredand that way we could stop having a pre-existing db, and instead start the db as part of a test fixture15:12
fungiand postgres has fine-grained permissions in that area already (we may even already set them?) which i believe allow us to grant the control we need safely. at least that's what i recall from prior discussions15:13
* mordred used to be core on drizzle15:13
mordredyup15:13
*** bauzas has joined #tripleo15:14
jog0so then short term lets bump up the timeout past 160?15:20
jog0mikal: ^ where is your make dbs faster 'turbo hipster' tool when we need it15:21
jog0err make sure we have no slow migrations15:21
*** lucasagomes_ has joined #tripleo15:22
*** matsuhashi has quit IRC15:23
jog0rpodolyaka: sounds like we have along term solution, just need a short term one too15:23
*** lucasagomes has quit IRC15:24
rpodolyakajog0: yeah, and OS_TEST_TIMEOUT seems to be the only one. I wish there was a way to increase it for specific tests15:24
jog0lifeless: ^15:24
jog0rpodolyaka: push a patch up to change, and say that you ideally want a per test change and hope someone knows15:25
jog0such as lifeless15:25
rpodolyaka jog0: ack15:25
jog0            self.useFixture(fixtures.Timeout(test_timeout, gentle=True))15:25
jog0I wonder if setting a different value will override that15:26
*** lucasagomes_ is now known as lucasagomes15:26
*** jcooley_ has joined #tripleo15:30
openstackgerritJames Slagle proposed a change to openstack/tripleo-image-elements: Add diskimage-builder element.  https://review.openstack.org/5936115:32
SpamapSflwang: auto, not just yet, but scaling, yes.15:39
flwangSpamapS: got, so do you think it's a good food to try?15:39
*** john-n-seattle1 has joined #tripleo15:42
SpamapSflwang: the problem with auto scaling is that you cannot decide which machines to remove.15:43
flwangSpamapS: you mean maybe there are something are running on it?15:44
*** john-n-seattle1 has quit IRC15:46
*** rongze has quit IRC15:46
SpamapSflwang: right. It is great for things like API nodes, but not so much for compute nodes. :)15:50
flwangSpamapS: yep, but actually, what I want to do is autoscaling the OpenStack cloud controller, such as the api services, mq, db, etc15:51
*** max_lobur is now known as max_lobur_afk15:51
SpamapSflwang: mq and db are also problematic15:52
SpamapSflwang: if you are in a degraded state already in your HA db/mq setup, you must not scale down until you have recovered.15:52
*** jistr|mtg has quit IRC15:53
*** vkozhukalov has quit IRC15:53
flwangSpamapS: good point, so how about the api services?15:53
SpamapSflwang: I think Heat autoscaling should be able to automate these cases, but it definitely does not today.15:53
SpamapSflwang: anything stateless is crazy easy. :)15:54
flwangSpamapS: yep, I will leverage Heat as well15:54
flwangSpamapS: yep, state is the key point15:54
flwangSpamapS: so overall, do you think it's worthy to try?15:55
SpamapSflwang: if you have need to scale out the API services, I'd happily accept a set of tripleo-heat-templates which make that possible.15:56
SpamapSflwang: for the other pieces, Heat is undergoing a lot of autoscaling changes that will perhaps make this possible later on.15:57
flwangSpamapS: cool, thanks for all your valuable suggestion15:57
flwangSpamapS: really helpful15:57
*** cd-undercloud has joined #tripleo15:58
cd-undercloud************** overcloud complete status=0 ************15:58
*** cd-undercloud has quit IRC15:58
openstackgerritYuriy Zveryanskyy proposed a change to openstack/diskimage-builder: Add deploy ramdisk element for Ironic  https://review.openstack.org/5977015:59
*** jistr has joined #tripleo15:59
*** jistr is now known as jistr|mtg16:00
*** rongze has joined #tripleo16:02
*** john-n-seattle1 has joined #tripleo16:03
*** funzo_ has joined #tripleo16:08
*** funzo has quit IRC16:08
*** rdopieralski has quit IRC16:16
*** jcoufal has quit IRC16:16
*** rbrady has quit IRC16:16
*** funzo_ has quit IRC16:17
mikaljog0: so, I am hoping to turn turbo hipster on today16:17
mikaljog0: just fighting final config stuff16:17
mikaljog0: but I have been abandoned to do it all myself, and I'm not very smart16:18
*** funzo has joined #tripleo16:20
*** rbrady has joined #tripleo16:26
*** rushiagr has joined #tripleo16:32
*** UtahDave has joined #tripleo16:35
jog0mikal: cool, I ask because we are seeing the py27 nova tests failing due to slow migration tests16:37
*** CaptTofu has joined #tripleo16:39
mikaljog0: ?16:40
mikaljog0: as in the unit test times out during a snake walk?16:40
SpamapSyou know16:41
SpamapSTripleO deployment has gotten slower of late16:41
jog0https://bugs.launchpad.net/openstack-ci/+bug/1216851 + scroll back16:41
uvirtbotLaunchpad bug 1216851 in nova "nova unit tests occasionally fail migration tests for mysql and postgres" [Undecided,Confirmed]16:41
SpamapSperhaps because of slow migrations (with empty databases no less!)16:41
jog0mikal: so yes snake walk16:41
SpamapSI wonder if we could make a job which automates the roll up.16:42
jog0SpamapS: maybe we can blame the new version of sqlalchemy-migrate that we released16:42
jog0maybe its now just slower for some reason16:42
mikaljog0: so that's going to be systemic. Its not a simple fix.16:42
jog0SpamapS: shouldn't be too hard to confirm16:42
SpamapSjog0: the migrations only got a little slower.. I don't have instrumentation to confirm when they got longer or how much longer they got.16:42
SpamapSjog0: though they've always been slower than they need to be.16:43
jog0mikal: yeah I know :( unless there is an easy explanation like new sqlalchemy-migrate (FWIW I am just hoping, I don't actually thing that is the problem)16:43
mikaljog0: not that I am aware of, but I am swamped because I'm travelleing and therefore in meetings all the time16:43
*** jcooley_ has quit IRC16:43
* jog0 wanders off back to the Hanukkah party 16:44
jog0anyway we have a short term workaround . bump up the timeout then we can dig into this without it being a gate issue16:44
mikalOk, cool16:45
mikalI do have a list of slow migrations I cn share16:45
mikalSo that might be a good starting point16:45
*** UtahDave has quit IRC16:47
*** shakayumi has joined #tripleo16:48
*** michchap_ has quit IRC16:48
*** cd-undercloud has joined #tripleo16:49
cd-undercloud************** overcloud complete status=0 ************16:49
*** cd-undercloud has quit IRC16:49
*** marun has quit IRC16:50
*** cwolferh_ has joined #tripleo16:52
*** jistr|mtg is now known as jistr16:54
*** cwolferh has quit IRC16:55
*** LinuxJedi_ is now known as LinuxJedi16:59
*** akuznetsov has quit IRC17:05
*** rongze has quit IRC17:05
*** rongze has joined #tripleo17:07
*** jcooley_ has joined #tripleo17:08
*** jcooley_ has quit IRC17:15
openstackgerritMark McLoughlin proposed a change to openstack/diskimage-builder: source-repositories: log the repo we're cloning from  https://review.openstack.org/6003517:19
*** jistr has quit IRC17:19
*** rushiagr has quit IRC17:22
*** rushiagr has joined #tripleo17:22
*** jdob_ has joined #tripleo17:26
*** pblaho has quit IRC17:27
*** jdob has quit IRC17:27
*** martyntaylor has quit IRC17:36
*** jdob_ has quit IRC17:38
*** jdob has joined #tripleo17:38
*** vkozhukalov has joined #tripleo17:39
*** matty_dubs is now known as matty_dubs|lunch17:42
*** cd-undercloud has joined #tripleo17:47
cd-undercloud************** overcloud complete status=42 ************17:47
*** cd-undercloud has quit IRC17:47
openstackgerritA change was merged to openstack/tripleo-image-elements: Add diskimage-builder element.  https://review.openstack.org/5936117:48
*** akuznetsov has joined #tripleo17:50
*** derekh has quit IRC17:50
*** michchap has joined #tripleo17:56
*** martyntaylor has joined #tripleo18:03
*** michchap has quit IRC18:03
*** akuznetsov has quit IRC18:03
* Ng dinners18:06
*** 16WABNAJZ has joined #tripleo18:11
*** 16WABNAJZ has quit IRC18:17
*** akrivoka has quit IRC18:19
*** lucasagomes has quit IRC18:25
* SpamapS polishes off a little utility to test the metadata + moustache templates embedded in elements18:31
*** matty_dubs|lunch is now known as matty_dubs18:31
*** athomas has quit IRC18:31
*** rbrady has quit IRC18:33
*** markmc has quit IRC18:34
lifelessmorning18:38
lifelessNg: hey so18:38
lifelessNg: what do you think of us capturing the build environment to a separate .build file?18:39
*** akuznetsov has joined #tripleo18:39
*** cd-undercloud has joined #tripleo18:40
cd-undercloud************** overcloud complete status=0 ************18:40
*** cd-undercloud has quit IRC18:40
openstackgerritA change was merged to openstack/diskimage-builder: Add deploy ramdisk element for Ironic  https://review.openstack.org/5977018:40
*** jprovazn has quit IRC18:42
*** jprovazn has joined #tripleo18:44
*** rushiagr has quit IRC18:45
openstackgerritJames Slagle proposed a change to openstack/diskimage-builder: Default name for ramdisks to deploy.  https://review.openstack.org/6004618:46
openstackgerritClint "SpamapS" Byrum proposed a change to openstack/os-apply-config: Log parsing errors from pystache as errors  https://review.openstack.org/6004718:47
dprinceslagle: you beat me to approving that ironic commit which has an issue for RHEL/Fedora18:49
* dprince volunteers slagle to fix it now :)18:50
dprinceslagle: see here https://review.openstack.org/#/c/59770/18:51
*** krotscheck has joined #tripleo18:55
slagledprince: see my reply :)18:55
*** michchap has joined #tripleo18:56
dprinceslagle: ah. We're good then.18:59
dprinceslagle: That busybox context always gets me (and has many times)18:59
dprinceslagle: I'm always like... this runs fine for me. Why won't busybox run it!18:59
slagleyea :)19:00
SpamapSwow, go us.. we have not created any invalid yaml or json in README.md in tripleo-image-elements19:00
*** michchap has quit IRC19:01
dprinceSpamapS: hi, question for you on https://review.openstack.org/#/c/59297/19:03
dprinceSpamapS: do you want me to go on and implement your upstart suggestions? Or would you prefer to do them?19:04
SpamapSdprince: if you are comfortable implementing it, I think I spelled it out well enough..19:04
SpamapSdprince: totally happy for somebody else to do it :)19:04
dprinceSpamapS: and we believe this will make the lifeless -2 go away19:04
dprinceSpamapS: well, not on this patch... but later in the series he -2'd me. His comments are on the etherpad here https://etherpad.openstack.org/p/tripleo-late-start-services19:05
* dprince just wants everyone to weigh in before he does the work19:07
SpamapSdprince: right we discussed a bit yesterday and then I thought about it.19:08
dprinceSpamapS: okay. The approaches seemed similar. I wasn't sure you both had talked...19:08
SpamapSdprince: upstart was kind of my thing at Canonical, so I am pretty confident this will work. I've wanted this built in to upstart since forever actually. :)19:08
dprinceSpamapS: On the systemd side I actually think my existing approach is sufficient.19:09
SpamapSdprince: agreed19:09
SpamapSbecause systemd has a sane enable/disable mechanism IIRC19:09
dprinceSpamapS: cool. Let me see if I can whip this up then.19:10
openstackgerritA change was merged to openstack-infra/tripleo-ci: Add element to install a testenv worker  https://review.openstack.org/5830519:24
*** rpodolyaka1 has joined #tripleo19:24
openstackgerritDan Prince proposed a change to openstack-infra/tripleo-ci: Drop notcompute element.  https://review.openstack.org/5911219:25
openstackgerritA change was merged to openstack-infra/tripleo-ci: Add script to install a testenv client  https://review.openstack.org/5830619:25
*** vkozhukalov has quit IRC19:26
openstackgerritClint "SpamapS" Byrum proposed a change to openstack/tripleo-image-elements: Create tests for example metadata and templates  https://review.openstack.org/6005419:27
openstackgerritClint "SpamapS" Byrum proposed a change to openstack/tripleo-image-elements: Add script to test all example metadata  https://review.openstack.org/6005519:27
openstackgerritClint "SpamapS" Byrum proposed a change to openstack/tripleo-image-elements: Run tests on elements in tox.ini  https://review.openstack.org/6005619:27
openstackgerritClint "SpamapS" Byrum proposed a change to openstack/tripleo-image-elements: Create tests for example metadata and templates  https://review.openstack.org/6005419:33
openstackgerritClint "SpamapS" Byrum proposed a change to openstack/tripleo-image-elements: Add script to test all example metadata  https://review.openstack.org/6005519:33
openstackgerritClint "SpamapS" Byrum proposed a change to openstack/tripleo-image-elements: Run tests on elements in tox.ini  https://review.openstack.org/6005619:33
*** rbrady has joined #tripleo19:39
openstackgerritJames Slagle proposed a change to openstack/diskimage-builder: Default name for ramdisks to image.  https://review.openstack.org/6004619:42
*** bauzas has quit IRC19:48
*** spzala has joined #tripleo19:49
openstackgerritDan Prince proposed a change to openstack/tripleo-incubator: Use set-os-type to set default NODE_DIST  https://review.openstack.org/5842619:56
*** michchap has joined #tripleo19:57
*** jdob has quit IRC20:00
*** michchap has quit IRC20:02
*** jdob has joined #tripleo20:03
*** boris-42 has quit IRC20:04
*** bauzas has joined #tripleo20:06
*** derekh has joined #tripleo20:10
lifelessok-> town for this thing, bbia few hours20:11
dprincepleia2/derekh: you guys ready to chat then?20:16
derekhdprince: ready20:16
pleia2dprince: yep20:16
*** boris-42 has joined #tripleo20:22
*** jtomasek has quit IRC20:28
openstackgerritJames Slagle proposed a change to openstack/diskimage-builder: Default name for ramdisks to image.  https://review.openstack.org/6004620:31
*** derekh has quit IRC20:46
dprincessh: connect to host cd-undercloud.tripleo.org port 22: Connection timed out20:47
dprinceCan anyone else connect to the CD environment?20:47
*** dprince has quit IRC20:50
*** vipul is now known as vipul-away20:51
Ngd0ugal: hrm, no20:52
SpamapSit's down for me too20:52
Ngon it via serial, checking for anything obviously full of fail20:53
Ngit doesn't immediately be able to ping anything on the public or private IP ranges20:53
SpamapSit may be mellanox fail again20:53
Ngdmesg is full of disk errors :/20:53
SpamapSNg: ignore loop disk errors20:54
Nglots of sdb20:54
SpamapSlovely20:54
Ngaaaaaand mellanox cmd_pending failures20:54
NgSpamapS: can we just rmmod/modprobe the mlx4 KOs? or is this a reboot?20:55
Ngmeh, rmmod doesn't actually work21:01
*** Ng changes topic to "CRITICAL: undercloud mlx4 driver failed again. Using OpenStack to deploy OpenStack; meetings Tuesday 1900 UTC in #openstack-meeting-alt"21:01
openstackgerritA change was merged to openstack/tripleo-incubator: Use set-os-type to set default NODE_DIST  https://review.openstack.org/5842621:01
Ngrather than just reboot, I'll hold things as they are, for someone to be around, who knows how to restart this without rebooting21:02
NgI don't want to put us back to square one on the rebooting thing, having only just dug ourselves out of that ;)21:02
*** morazi has quit IRC21:06
*** sballe has joined #tripleo21:15
*** noslzzp has quit IRC21:15
openstackgerritChris Jones proposed a change to openstack/tripleo-incubator: Assert undercloud auth before building overcloud.  https://review.openstack.org/5923721:16
*** jprovazn has quit IRC21:17
SpamapSNg: the rmmod often locks up for a while21:18
SpamapSNg: note that I found a newer version of the mlx driver .. haven't tried it yet though.21:19
Ngk, I'll leave it running for a bit, but I'm not going to be around for very much longer21:19
*** epim has joined #tripleo21:20
*** morazi has joined #tripleo21:24
*** julim has quit IRC21:25
*** vipul-away is now known as vipul21:25
SpamapSNg: you on via conman?21:25
SpamapSNg: I'll take over.21:26
openstackgerritClint "SpamapS" Byrum proposed a change to openstack/tripleo-image-elements: Add an ntp element  https://review.openstack.org/6007921:26
NgSpamapS: yeah I just logged in via conman (0017) and it refused to rmmod at all21:26
Ngoh, I wonder if my previous process is still running from before I logged out21:26
openstackgerritClint "SpamapS" Byrum proposed a change to openstack/tripleo-image-elements: Add script to test all example metadata  https://review.openstack.org/6005521:30
openstackgerritClint "SpamapS" Byrum proposed a change to openstack/tripleo-image-elements: Create tests for example metadata and templates  https://review.openstack.org/6005421:33
openstackgerritClint "SpamapS" Byrum proposed a change to openstack/tripleo-image-elements: Add script to test all example metadata  https://review.openstack.org/6005521:33
openstackgerritClint "SpamapS" Byrum proposed a change to openstack/tripleo-image-elements: Run tests on elements in tox.ini  https://review.openstack.org/6005621:34
jdobwhen does the bot post about a change, each patch set on a review?21:35
rpodolyaka1yep21:37
rpodolyaka1and when a change is merged21:37
jdobgotcha. threw me off when I kept seeing the same review mentioned multiple times21:38
*** spzala has quit IRC21:43
*** rbrady has quit IRC21:46
*** lsmola has quit IRC21:47
*** rongze has quit IRC21:49
*** CaptTofu has quit IRC21:53
*** rbrady has joined #tripleo21:55
*** rbrady has left #tripleo21:55
*** noslzzp has joined #tripleo22:10
*** shakayumi has quit IRC22:12
*** jdob has quit IRC22:16
*** rongze has joined #tripleo22:19
SpamapSrmmod mlx4_en finally finished, now waiting for mlx4_core22:21
SpamapSback up22:23
*** panda has quit IRC22:24
*** rongze has quit IRC22:28
openstackgerritElizabeth Krumbach Joseph proposed a change to openstack/tripleo-incubator: Add Elizabeth Krumbach Joseph to tripleo-cd-admins  https://review.openstack.org/6009722:29
lifelesspleia2: hey; it's great you want to step up into that, but I think you need significantly more familiarity with the ops plumbing we have - e.g. be approaching core status in tripleo22:30
lifelesspleia2: it's not a hard requirement, but all the current admins are well down that path with the one exception being derekh - but he's core in toci already which has ~all the same plumbing22:31
pleia2lifeless: ok22:31
*** morazi has quit IRC22:32
pleia2dprince and derekh suggested it on the call today so we could get some stuff set up, but they also have other systems we can collaborate on22:32
pleia2I follow up with them tomorrow22:33
lifelesspleia2: so, from a devops perspective22:33
*** rpodolyaka1 has quit IRC22:33
lifelesspleia2: I want to see all the bring be automated to the extent that a contributor can say 'can a tripleo-cd admin redeploy the test environment heat stack on the undercloud' and have that DTRT22:34
* pleia2 nods22:34
lifelesspleia2: e.g. you should be able to test locally with a virt bm cloud, be confident the images build etc, and point $whoever at the right script in toci to run to build+deploy and be done with it.22:34
lifelesspleia2: I am sure there will be glitches we find, but if there is any manual handholding, we start building up in-person-X-head-state22:35
lifelesswhich is bad22:35
pleia2lifeless: sure, we were just trying to work out some of the networking bits that were a bit hard to spec out in the call22:35
pleia2there are certainly other ways to do that than playing with the real CD cloud :)22:35
lifelesspleia2: ack; I think you should be able to test all those locally with a emulated bm + devstack, or somesuch22:36
lifelessyah22:36
openstackgerritClint "SpamapS" Byrum proposed a change to openstack/tripleo-image-elements: Run tests on elements in tox.ini  https://review.openstack.org/6005622:38
*** morazi has joined #tripleo22:47
*** cd-undercloud has joined #tripleo22:48
cd-undercloud************** overcloud complete status=1 ************22:48
*** cd-undercloud has quit IRC22:48
*** bauzas has quit IRC22:55
*** rongze has joined #tripleo22:55
*** rongze has quit IRC23:00
*** bauzas has joined #tripleo23:00
*** noslzzp has quit IRC23:02
*** noslzzp has joined #tripleo23:04
*** noslzzp has quit IRC23:08
*** ccrouch has quit IRC23:10
*** noslzzp has joined #tripleo23:12
*** cd-undercloud has joined #tripleo23:13
cd-undercloud************** overcloud complete status=1 ************23:13
*** cd-undercloud has quit IRC23:13
SpamapSERROR: Giving up waiting for overcloud delete to complete after 120 attempts.23:14
SpamapSI wonder if we could get the IRC message to contain the line number of devtest_overcloud.sh where we failed.23:14
*** matty_dubs is now known as matty_dubs|gone23:18
*** vipul is now known as vipul-away23:21
SpamapSlifeless: how can we encode this "move eth2's address to br-ctlplane" thing so we don't have to remember to do it?23:22
lifelessSpamapS: it is encoded23:22
lifelessSpamapS: idempotently even23:22
lifelessSpamapS: but we're not pushing the route from the seed atm, remember ?23:23
SpamapShow so?23:23
lifelessSpamapS: see the ovs agent script23:23
lifelessSpamapS: there is clearly a bug23:23
SpamapSlifeless: we had another mellanox freakout .. :-/23:24
lifelessSpamapS: oh, you mean with the mellanox workaround?23:24
lifelessSpamapS: NFI23:24
lifelessI just have a command line with ;'s between it23:24
SpamapSlike I want to say "make it the way it should be"23:24
*** michchap has joined #tripleo23:24
lifelessyeah, I'm with you on that23:24
SpamapSlifeless: Keep meaning to get back to building an image with the newer version of the driver.23:25
openstackgerritA change was merged to openstack-infra/tripleo-ci: Drop notcompute element.  https://review.openstack.org/5911223:33
openstackgerritA change was merged to openstack/tripleo-incubator: Assert undercloud auth before building overcloud.  https://review.openstack.org/5923723:34
openstackgerritA change was merged to openstack/diskimage-builder: Default name for ramdisks to image.  https://review.openstack.org/6004623:37
*** vipul-away is now known as vipul23:41
*** rongze has joined #tripleo23:56

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