Thursday, 2016-03-17

*** mihgen has quit IRC00:06
*** flwang has quit IRC00:07
*** dims has joined #openstack-release00:08
*** flwang1 has joined #openstack-release00:09
*** dims_ has quit IRC00:09
*** dims has quit IRC00:10
*** mihgen has joined #openstack-release00:14
*** jokke_ has quit IRC00:16
*** sridhar_ram has quit IRC00:16
*** jokke_ has joined #openstack-release00:16
openstackgerritMerged openstack-infra/release-tools: add convenience script for finding unreleased stable branch changes
*** flwang1 has quit IRC00:23
*** flwang1 has joined #openstack-release00:26
*** dims has joined #openstack-release00:52
*** njohnston has quit IRC01:04
*** kzaitsev_mb has quit IRC01:04
*** odyssey4me has quit IRC01:04
*** odyssey4me has joined #openstack-release01:06
*** njohnston has joined #openstack-release01:06
*** sdake has quit IRC01:22
*** kzaitsev_mb has joined #openstack-release02:00
*** odyssey4me has quit IRC02:19
*** mriedem1 has quit IRC02:19
*** mattoliverau has quit IRC02:19
*** sileht has quit IRC02:19
*** mtreinish has quit IRC02:19
*** tsufiev has quit IRC02:19
*** dansmith has quit IRC02:19
*** dims has quit IRC02:20
*** Qiming has joined #openstack-release02:23
*** odyssey4me has joined #openstack-release02:25
*** mriedem1 has joined #openstack-release02:25
*** mattoliverau has joined #openstack-release02:25
*** sileht has joined #openstack-release02:25
*** mtreinish has joined #openstack-release02:25
*** tsufiev has joined #openstack-release02:25
*** dansmith has joined #openstack-release02:25
*** johnsom_ has joined #openstack-release02:28
*** johnsom has quit IRC02:28
*** kzaitsev_mb has quit IRC02:28
openstackgerritSean McGinnis proposed openstack/releases: Tag Cinder Mitaka RC1
*** Qiming has quit IRC02:57
*** armax has joined #openstack-release03:09
armaxttx: I am gonna need an RC2 due to unforeseen bug 155820503:12
openstackbug 1558205 in neutron "neutron-lbaas plugin installation breaks on gates" [Critical,New] - Assigned to Brandon Logan (brandon-logan)03:12
*** kzaitsev_mb has joined #openstack-release03:25
*** kzaitsev_mb has quit IRC03:31
*** mriedem1 has quit IRC03:40
*** johnsom_ is now known as johnsom03:59
*** kzaitsev_mb has joined #openstack-release04:27
*** kzaitsev_mb has quit IRC04:32
*** armax has quit IRC04:33
*** armax has joined #openstack-release04:59
*** sdake has joined #openstack-release05:44
*** sdake_ has joined #openstack-release05:45
*** sdake has quit IRC05:49
*** sdake_ is now known as sdake05:51
*** kzaitsev_mb has joined #openstack-release06:29
*** kzaitsev_mb has quit IRC06:35
*** sridhar_ram has joined #openstack-release07:23
*** kzaitsev_mb has joined #openstack-release07:30
openstackgerritAndreas Jaeger proposed openstack/releases: openstackdocstheme 1.3.0 "Mitaka" release
*** AJaeger has joined #openstack-release07:32
AJaegerdhellmann, dims: We're ready with openstackdocstheme, please get this in - and also the corresponding change for requirements to update upper-constraints.07:32
*** kzaitsev_mb has quit IRC07:35
*** cody-somerville has quit IRC08:01
*** amrith is now known as _amrith_08:10
ttxarmax: probably better to let some time pass over RC1, unless the issue renders neutron totally untestable08:11
ttxi.e. give people a chance to raise other release-critical bugs, and not train QA/testers/packagers to ignore the RCs as being too ephemeral08:12
ttxusually I recommend at least 2 full days between each RC08:12
ttx(unless the current RC is unusable)08:12
ttxprocessing cinder08:16
*** sridhar_ram1 has joined #openstack-release08:17
*** sridhar_ram has quit IRC08:19
openstackgerritLingxian Kong proposed openstack/releases: Release Mistral rc1 for Mitaka
*** pcaruana has joined #openstack-release08:24
openstackgerritMerged openstack/releases: Tag Cinder Mitaka RC1
*** kzaitsev_mb has joined #openstack-release08:31
*** kzaitsev_mb has quit IRC08:36
*** johnsom has quit IRC08:40
*** johnsom has joined #openstack-release08:41
ttxkong: processing Mistral now08:42
kongttx: appreciate08:42
*** sdake has quit IRC08:43
*** sdake has joined #openstack-release08:44
openstackgerritMerged openstack/releases: Release Mistral rc1 for Mitaka
*** kzaitsev_mb has joined #openstack-release08:48
*** sdake has quit IRC08:50
armaxttx: yeah…I wasn’t suggesting to cut RC2 right way08:50
*** cody-somerville has joined #openstack-release08:52
*** cody-somerville has quit IRC08:52
*** cody-somerville has joined #openstack-release08:52
armaxttx: RC1 is usable, the issue lies with upstream CI, my only concern is that packagers may end up using the same testing infrastructure to gate their packaging, in which case they’d be broken as much as we were08:53
armaxfwiw we have known fixes and we’re back to sanity in both master and stable/mitaka08:53
ttxkong: all set, I'll let you announce it to mistral folks08:53
kongttx: ok, got it, thanks08:54
ttxkong: hmm wait08:54
ttxsome tarballs are still building08:54
ttxshould be there in 15 seconds though08:55
ttxso probably safe to announce anyway08:55
ttxkong: yep all set08:56
kongttx: :-)08:56
*** cody-somerville has quit IRC09:01
*** cody-somerville has joined #openstack-release09:02
skraynevttx: hi. it's again me ;) could you please help me understand how should be processed rc-2 ? now we have rc1 tag and stable branch. so we should not merge any code unrelated with rc-2 to master branch? or visa-versa - we can land any patches in master and backport necessary for rc-2 to stable/mitaka?09:04
ttxskraynev: if you find a bug that you'd like to include a fix for in mitaka, you would first fix it in master09:06
ttxthen propose a backport to stabel/mitaka09:06
*** onovy has quit IRC09:07
ttxonce all known release-critical issues are fixed in stable/mitaka, ask for another RC09:07
ttxIn master you can land anything09:07
ttxIn stable/mitaka you can only land bugfix backports09:07
skraynevttx: got it, so rc-2 will be based on stable/liberty branch09:07
ttxstable/mitaka, yes09:07
skraynevgot it. 1. fix in master 2. backport to stable/liberty. 3. ask rc2 when all necessary fixes will be there09:08
ttxstable/mitaka, not stable/liberty. Otherwise yes09:08
skraynevttx: thank you for the help :)09:08
*** kzaitsev_mb has quit IRC09:29
*** onovy has joined #openstack-release09:31
*** sridhar_ram1 has quit IRC09:46
*** sdague has joined #openstack-release09:51
*** dims has joined #openstack-release09:57
*** daemontool has joined #openstack-release10:14
*** kzaitsev_mb has joined #openstack-release10:29
*** daemontool has quit IRC10:31
*** kzaitsev_mb has quit IRC10:34
*** kzaitsev_mb has joined #openstack-release10:57
*** kzaitsev_mb has quit IRC11:02
openstackgerritDmitry Tantsur proposed openstack/releases: Release ironic-inspector 3.2.0 for Mitaka
dtantsurdhellmann, mitaka final when you have some time ^^11:37
*** kzaitsev_mb has joined #openstack-release11:39
*** openstackgerrit has quit IRC11:48
*** openstackgerrit has joined #openstack-release11:48
*** _amrith_ is now known as amrith12:18
*** gordc has joined #openstack-release12:18
*** armax has quit IRC12:20
gordcdhellmann: just an fyi re: ceilometer-rc1. i'm waiting on gate to be fixed so i can merge missing releasenote. should be better with this:
patchbotgordc: patch 293942 - openstack-infra/project-config - add uuid-runtime to base image builds12:25
*** dims has quit IRC12:25
*** armax has joined #openstack-release12:43
dhellmannAJaeger : did you come up with an approach that won't require a last-minute release for newton?12:43
dhellmanndtantsur : ack12:45
dhellmanngordc : ack12:45
*** amrith is now known as _amrith_12:55
*** amotoki has joined #openstack-release12:57
AJaegerdhellmann: YEs,changes 293858 and  29385913:13
AJaegerdhellmann: those remove the version dependent information13:13
*** bdemers has quit IRC13:19
dhellmannAJaeger : great, thanks13:23
dhellmannttx, dims: I'm going to process the openstackdocstheme release13:23
*** mriedem has joined #openstack-release13:24
AJaegerthanks, dhellmann13:26
openstackgerritMerged openstack/releases: openstackdocstheme 1.3.0 "Mitaka" release
*** _amrith_ is now known as amrith13:31
dhellmannttx, dims: I'm done and relinquish the baton13:34
AJaegerdhellmann: I rechecked 293859, please approve the constraints change as well.13:34
AJaegeroh, you did - thanks!13:34
dhellmannAJaeger : that's all part of the release process :-)13:34
AJaegerdhellmann: we did it at the same time ;)13:34
AJaegerdhellmann: Hope that's it for documentation and we don't run into any unforeseen problems...13:36
smcginnisttx: Trying to approve your gitreview update on stable/mitaka but doesn't look like I have +A nor +2 abilities.13:38
patchbotsmcginnis: patch 293879 - cinder (stable/mitaka) - Update .gitreview for stable/mitaka13:38
*** sigmavirus24_awa is now known as sigmavirus2413:47
ttxsmcginnis: hmm, you should have access as ptl let me check13:48
ttxsmcginnis: looks like you're missing from,members13:49
ttxI'll take the liberty to fix that13:49
smcginnisttx: Huh, looks like I never made it on that one. Thanks!13:49
ttxsmcginnis: should be fine now. The pre-release branch is under release management + cinder-milestone control, to havind random patch landing13:50
ttxso requires us (release or ptl) to approve it13:50
smcginnisttx: Logged out and back in and still only -1/+1. I'll wait a bit and try again.13:51
ttxmaybe there is some lag, otherwise ask infra13:51
ttxdid all I could :)13:51
ttxanyway, I expect release team to pick it up and approve it13:52
ttxso no big deal13:52
smcginnisttx: Yeah, I'm sure it's just propogating or something. Thanks for your help! :)13:52
*** dims has joined #openstack-release13:53
*** bdemers has joined #openstack-release13:56
jrollsmcginnis: hard refresh should do it iirc14:10
smcginnisjroll: Open/close browser?14:11
jrollsmcginnis: ctrl-shift-r (or cmd-shift-r on os x)14:12
smcginnisjroll: Nope, same thing. But thanks, I learned something new!14:12
smcginnisjroll: Yeah, even tried switch from Chrome to Firefox. Same thing.14:13
dhellmannttx: dropping offline for my commute, bbiab14:18
*** dims has quit IRC14:19
*** amrith is now known as _amrith_14:19
*** dims has joined #openstack-release14:22
smcginnisjroll: And now it works. Go figure.14:25
*** dims has quit IRC14:25
smcginnisttx: Did you see the comments on here:
patchbotsmcginnis: patch 293880 - cinder - Update reno for stable/mitaka14:25
*** pcaruana has quit IRC14:28
*** _amrith_ is now known as amrith14:32
*** sdake has joined #openstack-release14:39
*** sigmavirus24 is now known as sigmavirus24_awa14:40
*** sigmavirus24_awa is now known as sigmavirus2414:41
ttxsmcginnis: I asked dhellmann about it and he recommended keeping mitaka last. I can ask again14:41
smcginnisttx: Hmm, seems odd. I would be interested to hear from dhellmann why.14:45
smcginnisttx: But fair enough, if that's what it will be everywhere.14:45
*** bdemers has quit IRC14:57
*** TravT has joined #openstack-release15:18
*** bdemers has joined #openstack-release15:24
mriedemttx: smcginnis: it was generated via a script,15:28
mriedemi re-ordered those in the nova change15:28
smcginnismriedem: So there isn't a reason for wanting it the way it is now?15:28
smcginnismriedem: Good - thanks!15:29
*** amrith is now known as _amrith_15:43
skraynevttx: I again need your help with launchpad magic. where can I configure rights for +2 and +a for new stable mitaka branch ? currently I can do only +-1 , (for other stable branch I have ability to +2 and +A)15:44
ttxskraynev: that would be because you're not in,members15:45
ttxLet me fix that15:45
ttxskraynev: that should work now, after a small wait15:46
ttxNote that you should only approve release-critical backports if they also landed in master. You can also just let release managers handle that under after release15:46
AJaegerdhellmann, dims, ttx: Tomorrow morning first translation proposals for stable/mitaka should show up ...15:47
*** sridhar_ram has joined #openstack-release15:49
skraynevttx: heh. we have one couple Medium and couple High bugs which we planned to land in rc2... do we need re-review their statuses ? Ideally will be good to backport them in mitaka with current statuses. Fixes are not Critical, but important..15:49
ttxAJaeger: good timing15:51
ttxskraynev: so you land them in master first, then propose backports.15:52
ttxThe important part is to check that the issue is fixed in master... otherwise you introduce a regression15:52
AJaegerttx, translation server was broken due to some openstackid changes ;( but it's up again now...15:52
skraynevttx: ok.15:53
*** pcaruana has joined #openstack-release16:22
*** doug-fish has quit IRC16:22
*** doug-fish has joined #openstack-release16:23
*** doug-fish has quit IRC16:27
*** doug-fish has joined #openstack-release16:38
*** doug-fish has quit IRC16:38
*** doug-fish has joined #openstack-release16:38
*** sigmavirus24 is now known as sigmavirus24_awa16:45
*** sigmavirus24_awa is now known as sigmavirus2416:45
*** doug-fish has quit IRC16:45
*** doug-fish has joined #openstack-release16:46
*** doug-fish has quit IRC16:46
ttxdhellmann: tags in repositories that are not in kilo/liberty/mitaka releases files:
*** doug-fish has joined #openstack-release16:46
ttxlots of false positives. My script can store things it should ignore and not report them next time16:47
ttxOnce we do the initial pass and I clean it up and add ignores, I'll probably submit it to release-tools16:47
ttxI'll etherpad it16:48
*** yarkot_ has joined #openstack-release16:48
*** mriedem1 has joined #openstack-release16:49
*** mriedem has quit IRC16:49
*** armax has quit IRC17:02
openstackgerritDmitry Tantsur proposed openstack/releases: Release ironic-inspector 2.2.5 for Liberty
dtantsurdhellmann, mriedem1, also a liberty release, if you have some time ^^^17:04
mriedem1will look17:05
*** mriedem1 is now known as mriedem17:05
dhellmannttx: I'm curious about the reno tags as false positives. Those releases were all made through the releases repo.17:15
ttxoh. Got it17:15
ttxit should scan _independent for release:independent stuff17:16
ttxconsier the results only valid for teh cycle stuff17:16
ttxwill do another pass tomorrow :)17:17
ttxlooks like we don't have any RC1 today17:18
ttxdid a couple this morning17:18
ttxMaybe mention the missing ones in your weekly reminder ?17:19
ttxif you haven't already17:19
*** yarkot_ has quit IRC17:19
* ttx catches up with email17:19
dhellmannI haven't sent that -- long commute this morning -- but I will17:23
* ttx disappears for evening17:30
*** kzaitsev_mb has quit IRC17:47
*** dims has joined #openstack-release17:50
*** sigmavirus24 is now known as sigmavirus24_awa18:00
*** doug-fish has quit IRC18:00
*** doug-fish has joined #openstack-release18:00
dimssorry was awol since AM, back now dhellmann and ttx18:03
*** doug-fish has quit IRC18:05
*** gordc has quit IRC18:07
*** pcaruana has quit IRC18:18
*** sigmavirus24_awa is now known as sigmavirus2418:28
*** pcaruana has joined #openstack-release18:30
*** sridhar_ram has quit IRC18:31
*** dtantsur is now known as dtantsur|afk18:36
*** gordc has joined #openstack-release18:44
*** doug-fish has joined #openstack-release18:45
*** doug-fish has quit IRC18:45
*** doug-fish has joined #openstack-release18:45
*** doug-fish has quit IRC18:50
*** doug-fish has joined #openstack-release18:51
dimsdhellmann : i see some kilo and liberty releases, should i process them? (oslo.db + oslo.messaging for kilo, glance-store and ironic-inspector for liberty)?18:56
dhellmanndims : I was going to hold off until mriedem reviewed those18:56
dhellmannso if he has, go for it18:57
dimsdhellmann : will check and ship those that have mriedem's ok18:58
dhellmanndims : ack, let me know when you're done. When I finish catching up on email, I'm going to process ironic-inspector18:59
mriedemdims: dhellmann: the oslo.db kilo one is mine18:59
*** sridhar_ram has joined #openstack-release18:59
mriedemwe need that to unblock ceilometer in kilo19:00
mriedemi +1ed the ironic-inspector for liberty one19:00
mriedemi had questions in the glance-store one19:00
mriedemand haven't seen the oslo.messaging one19:00
patchbotdims: patch 292602 - releases - oslo.messaging 1.8.4 - kilo19:00
dimsdhellmann : on it19:01
openstackgerritMerged openstack/releases: Release ironic-inspector 2.2.5 for Liberty
openstackgerritMerged openstack/releases: kilo: release oslo.db 1.7.4
dimsmriedem : do you fix u-c for those releases by hand?19:09
mriedemdims: which releases?19:10
dimsmriedem : any release we make say for liberty19:10
mriedemno i haven't been touching upper-constraints in stable/liberty19:11
mriedemb/c of the previous bot update snafus19:11
dimsdhellmann : those two release above have hit pypi. so done with those19:11
dimsmriedem : ack19:11
mriedemi'm checking this oslo.messaging kilo release request still19:11
mriedemlots of backports in oslo.messaging scares me19:12
dimsack mriedem i can do that after dhellmann releases the mitaka ironic-inspector19:12
mriedemi see at least one backport was messed up and had to be manually fixed later19:12
dhellmannmriedem : thanks19:12
dhellmanndims : I'll start that now19:12
dimsmriedem : all from redhat folks?19:13
mriedemdims: didn't really look19:13
mriedemi just know it's very hard to tell when something was released in oslo.messaging that causes a regression19:14
dimschanges proposed by John, Mehdi, Flavio
mriedemor that we're backporting something which introduced a regression and was later fixed but that fix wasn't backported19:14
mriedemyeah i know, i'm going through that19:14
dimsmriedem : yep, i know :(19:15
patchbotmriedem: patch 217770 - oslo.messaging (stable/kilo) - Fix gate jobs in stable/kilo (4 backports) (MERGED)19:15
mriedemthat bumps the minimum required version of pyngus for kilo on that release19:16
dimsmriedem :
mriedembut w/o it we couldn't land changes to oslo.messaging in kilo19:18
mriedemwell, e.g.
dimsmriedem : looking at the diffs above... the only code changes relevent are oslo_messaging/_drivers/ and oslo_messaging/_drivers/ which don't seem too bad19:20
mriedemso i guess we'll just know if the red hat guys are backporting the wrong things b/c they are obviously testing it downstream19:20
mriedemdims: there is still a new higher minimum dependency bump in here19:21
openstackgerritMerged openstack/releases: Release ironic-inspector 3.2.0 for Mitaka
dimsmriedem : i am ignoring any changes to proton and focusing only on the ones that hit rabbitmq19:22
dimsto see if anyone else will be affected19:22
mriedemwhat was the support statement on amqp1/proton in kilo?19:23
mriedemand it's gone now?19:23
dimswe still have amqp1/proton19:23
mriedemrather than backporting those changes to kilo, we probably should have just disabled the amqp1 job in kilo19:24
mriedemle sigh19:24
mriedemyeah i would have just skipped gate-oslo.messaging-dsvm-functional-amqp1-f21 for kilo19:26
mriedemonce it was busted and required requirement updates19:26
dimslesson learned19:26
mriedemdims: anyway, it's up to you and what you're comfortable releasing19:26
mriedemi don't think i'm going to +1 that one19:26
dhellmanndtantsur|afk, jroll : do you want a stable/mitaka branch for ironic-inspector at 3.2.0?19:27
jrolldhellmann: I believe so but dtantsur|afk is the right person to ask there19:28
dhellmannjroll : ok, I'll wait for confirmation19:28
dimsmriedem : -2'ed
mriedemdims: ok, i commented. i'd be +1 if the pyngus stuff was reverted, which would also mean skipping that amqp job on kilo, but i'm obviously ok with that19:31
mriedemdims: note that at the summit we're going to be talking about extending EOL for kilo19:31
mriedemby at least 6 months19:31
dimsmriedem : ack19:31
mriedembut that mostly applies to server projects19:31
mriedemdhellmann: which reminds me, i got thinking about the stable:follows-policy tag, i kind of think that should only be set on server projects,19:32
mriedemit's a bit weird to set that on clients/libraries19:32
mriedemi think we expect people could be git installing from stable branches for server projects, but not so much so for libs/clients19:32
mriedemwe really only have the stable branches for backporting necessary things in the clients/libs, sort of different reasons than for the servers imo19:33
dhellmannmriedem : I would need to understand how that interacts with lifeless' work on stable libs.19:33
dimsdhellmann : looking at spreadsheet and the email to -dev, we should probably be ok with cutting stable branches for requirements/devstack/grenade/devstack-gate i think. right?19:33
mriedemi don't think it does at this point19:33
dhellmanndims : I don't want to rush that19:34
dhellmannmriedem : ok19:34
dimsdhellmann : ack19:35
dhellmanndims : I think you're right, based on a minimal set of projects, though19:35
*** cody-somerville has quit IRC19:35
dhellmanndims : we know neutron has another rc planned19:36
dimsdhellmann : y, it's just that its going to take sometime to get the jobs up and running and green. would be great to have all green when we ship :)19:36
dimsack on neutron19:36
dhellmanndo we have a list of the changes we know we need to make those jobs work?19:37
*** doug-fish has quit IRC19:38
dims[06:57:27]  <sdague>dansmith and I were talking it through the other day19:38
dims[06:57:36]  <sdague>once requirements gets a branch, then we need to branch devstack19:38
dims[06:57:44]  <sdague>then grenade19:38
dims[06:57:54]  <sdague>devstack-gate does everything on the branch selection19:38
dimshad a chat with sdague in the AM19:38
dansmithI have tool for that!19:38
patchbotdansmith: patch 293673 - openstack-dev/devstack - Add tool for detecting releas...19:39
dimshey dansmith !19:39
dansmithyou can blame sdague for it not being merged yet :)19:39
dimsLOL dansmith19:39
dhellmanndansmith : nice19:43
*** onovy has quit IRC19:43
dansmithdhellmann: for example:19:44
dansmithUnable to branch yet. Some projects do not have stable branches for mitaka:19:44
dhellmanndansmith : the requirements repo isn't in the list?19:45
dansmithdhellmann: it's just projects in devstack.. sdague was supposed to get back to me about whether we needed to consider requirements for that branching go/no-go decision19:45
dansmithhe's off futzing with "fixing api docs" or some silly task :D19:46
dhellmannI think we have been assuming requirements did need to branch first to avoid having the requirements auto-update logic drag values from master into the stable tests19:46
dimsdansmith : when i talked to him this morning, see chat snippet above. he did say requirements go first19:46
dansmithdims: okay19:46
dhellmannbut I don't know where that logic lives off the top of my head (devstack or d-g)19:46
dansmithdhellmann: yeah, me either, but if we think it's a requirement I should just add it to that tool for safety19:47
*** cody-somerville has joined #openstack-release19:47
dims+1 dansmith19:47
notmynameFWIW, swift is planning on it's mitaka release next week, and there are no known dependency updates planned19:52
*** doug-fish has joined #openstack-release19:53
dhellmannnotmyname : thanks19:53
notmynamewell, hang on about dependency updates19:54
dimsdhellmann : dang, i had that cell ready :)19:54
notmynamethere are no planned dependency updates that aren't already on master19:54
dhellmannnotmyname : as long as what you have is compatible with the global list, it's fine19:55
notmynamethere are changes in swift's requorements.txt file since the last release19:55
notmynameroght, ok19:55
notmynameit is (otherwise wouldn't have landed)19:55
dhellmannwe're working out the sequence of when to branch requirements, devstack, etc. and swift is one of the blockers for that19:55
*** _amrith_ is now known as amrith19:55
dimsnotmyname : in general swift would work in an environment with latest g-r/u-c right?19:56
notmynameok. I think I'm trying to say that there's no technical reason for it to be a blocker at this point19:56
notmynamedims: yes19:56
dimscool. thanks notmyname19:56
notmynamedims: my understanding is that's what's tested in the gate19:56
dhellmannnotmyname : I think it has to do with the way the gate jobs are set up, esp. for grenade? dansmith, sdague do we actually need to block on swift?19:57
dansmithI think we do19:57
sdagueI'm pretty sure we need a swift branch, it's in the main job19:58
dansmithsdague was saying something about it yesterday to me and specifically mentioned swift I think19:58
sdaguewe could stop that linkage and pull swift out of it, we haven't done that as of yet19:58
dhellmannhow much work would that be?19:59
dhellmannand what are the downsides?19:59
sdaguenot a ton of work, we just need to flip some bits in d-g. The current coupled structure is pretty much legacy for some of those things, as the swift testing is mostly secondary in those tests via glance backend.20:01
sdaguejuggling a few other things, maybe we can discuss tomorrow20:01
dimssdague : do we need stuff like this to be fixed in grenade as well? (
sdaguedims: yes, there would be some work, I have a meeting right now then have to make dinner, so you're only going to get trickle bw out of me atm. But tomorrow would be good discussion.20:03
dims:) sdague20:03
*** doug-fish has quit IRC20:07
*** doug-fish has joined #openstack-release20:13
*** pcaruana has quit IRC20:17
*** kzaitsev_mb has joined #openstack-release20:21
*** prometheanfire has joined #openstack-release20:28
prometheanfireanyone mind taking a look at this upper-constraints bump for liberty?
patchbotprometheanfire: patch 291859 - requirements (stable/liberty) - Update libvirt-python upper-requirement20:29
prometheanfirebeen up a week20:29
*** pcaruana has joined #openstack-release20:30
dhellmannprometheanfire : can you add details to the patch explaining why we need the update? we're trying to keep the stable branches locked down as much as we can20:34
prometheanfirebecause some distros have updated their stable libvirt-python?20:36
dhellmannprometheanfire : please put it in the patch so the stable team can review it20:36
dhellmannprometheanfire : we're going to be much more cautious about changing stable constraints from this point forward, to avoid some of the issues we've had in the past with getting them out of sync. This sounds like a reasonable change, but...20:37
prometheanfireit's causing conflicts in gentoo at least20:39
prometheanfirecurrent cap is 1.3.020:39
prometheanfireour current stable is 1.3.120:40
prometheanfirewith 1.3.2 avail for install20:40
dhellmannok. we separate our test constraints from the actual dependencies of packages specifically to avoid this20:50
dhellmannthe idea is that you would try to package the constraints, but set up dependencies using the regular rules that allow overlapping ranges20:50
prometheanfirehow do I know the upper constriants?20:52
prometheanfireI don't want my packages constantly breaking20:52
prometheanfireI want it tested, convieniently openstack does this for me with upper-constraints20:53
dhellmannprometheanfire : upper-constraints is called that because we also plan a lower-constraints at some point, not because we consider them to be a cap20:59
prometheanfirethey do operate as a cap though20:59
prometheanfirewhatever you consider them20:59
prometheanfirethose are the max versions you test with21:00
prometheanfireI have nowhere else to get caps from, and as I said in the commit message I need that21:00
prometheanfireas we have multiple versions of openstack available21:01
dhellmannonly in some jobs, and we change them pretty regularly in master21:01
prometheanfireand I update my ebuilds as well :P21:01
dhellmannok, well, we'll review the update but you should expect the constraints on a stable branch to be pretty hard to change for the short term21:02
prometheanfireshould I expect all the oslo packages to work with liberty?21:03
prometheanfirethe versions that are for mitaka21:03
dhellmannI don't think we're testing master oslo libs with liberty, no21:03
dhellmannI don't think we've done anything to explicitly break them, but I'd have to check with dims21:04
prometheanfireso I need to pull my list of caps for oslo from somewhere21:04
dhellmannwe had some major version updates when we dropped py 2.621:04
prometheanfireno oslo caps in global-reqs or the nova's requirements.txt (as an example)21:04
dhellmannI don't know how many other distros are trying to support multiple versions of openstack in one package repo21:04
dhellmannyeah, we don't cap our requirements any more because that led to other issues21:05
prometheanfireit's a complex problem :D21:05
dhellmannI agree. I'm about to have to drop offline, maybe you want to start a ML thread on it?21:05
prometheanfireit's mainly an issue for packages outside of openstack though21:05
prometheanfireanother one?21:05
prometheanfireI did the one when it was first added to21:05
prometheanfirebut will do21:06
*** sdague has quit IRC21:06
dimsdhellmann : me, mriedem and lifeless were talking about that. at any given point master release of all oslo libs should work with the previous release. (theoretically we want to be able to do that)21:06
dhellmanndims : yes, but are we actually testing it yet?21:07
dimsdhellmann : we were until we stopped stable/liberty from accepting the bot updates for u-c21:07
dimswe still have src jobs against liberty21:07
prometheanfireit'd be nice if I could not cap21:07
prometheanfireit's a pain to update21:08
dhellmannwe should try to talk about that early in the week at summit. usually it comes up friday, and I'm too tired to comprehend the issues21:08
prometheanfireI'll be there as well21:08
dimsso we have the libs from stable/liberty that should work with stable/liberty and libs from mitaka that should work with stable/liberty as well. so 2 sets of possible libraries21:08
dims++ dhellmann21:08
prometheanfiredims: as long as the current releases libs work with the previous that's all I want21:09
prometheanfireit'll actually make it much easier to package21:09
lifelesso/ everyone21:09
dimshey lifeless21:09
lifelessprometheanfire: thats good to know, adds ammo to my spec on doing that testing21:09
prometheanfirehere's all the olso-messaging versions I support21:09
prometheanfire     Available versions:  ~1.4.1 ~1.5.1 ~1.7.0 ~1.8.0 1.8.2{tbz2} ~1.8.3 ~1.9.0 ~1.10.0 2.4.0{tbz2} {test PYTHON_TARGETS="python2_7 python3_3 python3_4"}21:09
lifelessprometheanfire: gentoo? or nixos?21:10
prometheanfireit really needs cleanup to be honest21:10
prometheanfirebut I'm cleaning out kilo when mitaka happens21:10
prometheanfireand if mitaka libs work in liberty... even better21:11
prometheanfirelifeless: what's that timeline on 'dual testing'?21:12
lifelessprometheanfire: mriedem and I don't yet see eye to eye on some of the basic assumptions21:13
lifelessprometheanfire: until we do, we don't have consensus21:13
lifelessprometheanfire: that said, I think everyone agreed that we should have tests that each branch of oslo works with the matching release of the servers21:14
lifelessprometheanfire: (probably as non-voting so we don't setup a wedge situation21:15
prometheanfireya, definitely non-voting21:15
lifelessprometheanfire: (but so we can tell before we cut releases, that things are ok)21:15
amrithdhellmann, quick update for you. we caught a bad break today. percona xtradbbackup released a new version (2.3.4) today that breaks our tests because of some additional validation. We're fixing that and will be in a position to call rc1 shortly thereafter.21:15
prometheanfiredo we have a room/time for this discussion at the conf?21:15
prometheanfireI'd like to be notified of it when it happens21:17
prometheanfirehaving upstream tell downstream what versions of dependencies are supported is 'standard' for packaging21:17
dhellmannamrith : ok, thanks21:17
prometheanfireatm that only exists in upper-constraints, which is a poor solution21:18
lifelessprometheanfire: uhm no21:18
lifelessprometheanfire: upper-constraints is *a* known good21:18
lifelessprometheanfire: its not the only good solution21:18
lifelessprometheanfire: the per package metadata documents known-bad interactions21:19
prometheanfireso I should be using that as version caps for at least the server packages (nova/neutron/etc)?21:19
prometheanfiredhellmann was telling me diferent I think21:20
lifelessThis is no worse than any C package using autoconf, for instance - and substantially better (since *a* known good is published)21:20
lifelessprometheanfire: using upper constraints for a release as exact matches is a safe thing to do21:20
prometheanfireI use what the package itself states as min constraints, and what upper-constraints says for caps21:21
dhellmannlifeless : it's safe, unless the result is conflicts because of using a specific === requirement instead of a range, which is what prometheanfire has now21:21
dhellmann(the conflict)21:21
prometheanfireya, within the openstack world it works21:22
prometheanfirebut at the edges (libvirt-python) it starts to fail unless updated21:22
*** gordc has quit IRC21:25
lifelessdhellmann: yes, lower + upper-as-=== is fail by definition21:27
lifelessdhellmann: I think we're on the same page21:27
dhellmannlifeless : ok. I think what prometheanfire is doing is using constraints as ===, or at least <=21:28
prometheanfirebut ya21:28
lifelessdhellmann: lower-from-packages + <= from u-c should be resolvable, and unlikely to fail21:28
*** sdake has quit IRC21:28
prometheanfireit has worked mostly good21:29
lifelessnot as safe as using u-c as-is, because of the possibility of undetected bad versions we haven't blacklisted21:29
prometheanfireexcept at the edges :P21:29
dhellmannlifeless : prometheanfire has both liberty and mitaka packages in the same package repo?21:29
dhellmannprometheanfire : maybe you can explain what is failing for you that requires us to update our libvirt in liberty21:29
*** amotoki has quit IRC21:29
dhellmannprometheanfire : I'm dropping off, sorry about the timing21:30
prometheanfiresince we can have multiple versions of a package in our tree we also mark one stable21:30
prometheanfirea newer version then what's in upper constraints was marked stable21:30
*** sdake has joined #openstack-release21:31
prometheanfirewhile the older (in upper constraints) will be eventually removed21:31
prometheanfireso right now it throws a warning21:31
prometheanfiresoon it will be an error21:31
openstackgerritSergey Lukjanov proposed openstack/releases: Sahara Mitaka RC1 release
lifelessprometheanfire: so we can't move libvirt versions in u-c easily21:34
lifelessprometheanfire: because libvirt (not pytohn-libvirt) comes from the os for us21:35
prometheanfireso how can it already be 1.3.2 in master?21:35
*** doug-fish has quit IRC21:35
lifelessprometheanfire: I may be wrong - but I thought python-libvirt had to match the C libvirt version21:36
lifelessprometheanfire: I think - haven't checked - that in CI we preinstall libvirt-python or some other special casing takes place.21:37
lifelessprometheanfire: worth digging into - anyhow, the point remains - u-c is *a* known good, not *the only* one21:37
prometheanfiredon't think it does21:38
prometheanfirethat's why they decoupled21:38
*** doug-fis_ has joined #openstack-release21:40
*** mriedem has quit IRC21:41
*** sdake has quit IRC21:45
*** doug-fis_ has quit IRC21:45
*** doug-fish has joined #openstack-release21:45
*** doug-fish has quit IRC21:50
*** doug-fish has joined #openstack-release21:56
*** amotoki has joined #openstack-release22:00
*** dims has quit IRC22:01
*** dims_ has joined #openstack-release22:01
*** amotoki has quit IRC22:17
*** amrith is now known as _amrith_22:19
*** dims has joined #openstack-release22:23
*** dims_ has quit IRC22:26
*** sigmavirus24 is now known as sigmavirus24_awa22:40
*** mriedem has joined #openstack-release22:54
*** sridhar_ram1 has joined #openstack-release23:01
*** sridhar_ram has quit IRC23:01
*** cody-somerville has quit IRC23:03
*** cody-somerville has joined #openstack-release23:04
*** dims has quit IRC23:08
*** TravT has quit IRC23:11
*** sdake has joined #openstack-release23:24
*** dims has joined #openstack-release23:28
*** dims has quit IRC23:32

Generated by 2.14.0 by Marius Gedminas - find it at!