Wednesday, 2017-02-08

« Tuesday, 2017-02-07 Index Thursday, 2017-02-09 »
*** bobh has joined #openstack-release00:07
*** sdake_ is now known as sdake00:10
*** hongbin has quit IRC00:24
*** cody-somerville_ has joined #openstack-release00:25
*** bobh has quit IRC00:35
*** mriedem has joined #openstack-release00:37
*** adrian_otto has joined #openstack-release00:38
*** david-lyle has quit IRC00:42
*** cody-somerville_ has quit IRC00:43
dhellmannmwhahaha : yeah, put the release note links in a second patch00:43
dhellmannwe don't want links to things that don't exist, so we have to do it in stages00:43
*** wolverineav has joined #openstack-release00:44
*** wolverineav has quit IRC00:48
*** zhurong has joined #openstack-release00:51
*** bobh has joined #openstack-release00:51
*** claudiub has quit IRC01:02
*** wolverineav has joined #openstack-release01:03
*** david-lyle has joined #openstack-release01:04
mwhahahadhellmann: I don't suppose this is documented anywhere01:10
*** wolverineav has quit IRC01:10
*** yarkot has quit IRC01:10
*** greenlight_ has left #openstack-release01:21
*** sdake has quit IRC01:24
*** wolverineav has joined #openstack-release01:25
*** yanyanhu has joined #openstack-release01:35
*** wolverineav has quit IRC01:44
*** gcb has joined #openstack-release01:47
*** wolverineav has joined #openstack-release01:50
*** wolverineav has quit IRC01:58
*** bobh has quit IRC02:00
*** adrian_otto has quit IRC02:04
*** amotoki has joined #openstack-release02:07
*** cody-somerville has quit IRC02:19
*** wolverineav has joined #openstack-release02:21
*** wolverineav has quit IRC02:26
*** adrian_otto has joined #openstack-release02:27
*** sdake has joined #openstack-release02:39
*** zhurong has quit IRC02:51
*** cody-somerville has joined #openstack-release02:55
*** cody-somerville has quit IRC02:55
*** cody-somerville has joined #openstack-release02:55
*** amotoki has quit IRC03:00
*** sdake has quit IRC03:01
*** sdake has joined #openstack-release03:02
openstackgerritJeremy Liu proposed openstack/releases master: Add a whitespace between words  https://review.openstack.org/43052403:05
*** mriedem has quit IRC03:07
*** amotoki has joined #openstack-release03:10
*** Rockyg has quit IRC03:12
*** bobh has joined #openstack-release03:13
*** amotoki has quit IRC03:15
*** bobh has quit IRC03:17
*** amotoki has joined #openstack-release03:19
*** zhurong has joined #openstack-release03:33
*** armax has quit IRC03:35
*** armax has joined #openstack-release03:36
*** armax has quit IRC03:36
*** amotoki has quit IRC03:40
*** hongbin has joined #openstack-release03:51
*** sdake_ has joined #openstack-release03:54
*** ricolin has joined #openstack-release03:55
*** sdake has quit IRC03:57
*** zhurong has quit IRC04:04
*** amotoki has joined #openstack-release04:11
*** udesale has joined #openstack-release04:16
*** wolverineav has joined #openstack-release04:23
*** askb has quit IRC04:24
*** hongbin has quit IRC04:27
*** wolverineav has quit IRC04:28
*** askb has joined #openstack-release04:30
*** csomerville has joined #openstack-release04:35
*** cody-somerville has quit IRC04:35
*** wolverineav has joined #openstack-release04:46
*** wolverineav has quit IRC04:50
*** amotoki has quit IRC05:03
*** amotoki has joined #openstack-release05:04
*** udesale__ has joined #openstack-release05:15
*** udesale has quit IRC05:18
*** csomerville has quit IRC05:20
*** wolverineav has joined #openstack-release05:55
*** wolverineav has quit IRC06:00
*** amotoki_ has joined #openstack-release06:13
*** adrian_otto has quit IRC06:13
*** amotoki has quit IRC06:14
*** zhurong has joined #openstack-release06:22
*** udesale has joined #openstack-release06:24
*** udesale__ has quit IRC06:26
*** sdake_ is now known as sdake06:27
*** claudiub has joined #openstack-release06:32
*** wolverineav has joined #openstack-release06:33
*** wolverineav has quit IRC06:38
*** wolverineav has joined #openstack-release06:39
*** cody-somerville_ has joined #openstack-release06:39
*** adrian_otto has joined #openstack-release06:43
*** zhurong has quit IRC06:43
*** wolverineav has quit IRC06:44
*** zhurong has joined #openstack-release06:49
*** kzaitsev_mb has quit IRC06:55
*** wolverineav has joined #openstack-release06:58
*** udesale__ has joined #openstack-release07:03
*** udesale__ has quit IRC07:04
*** wolverineav has quit IRC07:05
*** zhurong has quit IRC07:05
*** udesale__ has joined #openstack-release07:05
*** udesale has quit IRC07:05
*** udesale__ has quit IRC07:07
*** wolverineav has joined #openstack-release07:10
*** wolverineav has quit IRC07:15
*** wolverineav has joined #openstack-release07:15
*** wolverineav has quit IRC07:20
*** sdake_ has joined #openstack-release07:23
*** sdake has quit IRC07:26
*** pcaruana has joined #openstack-release07:29
*** adrian_otto has quit IRC07:33
*** david-lyle_ has joined #openstack-release07:37
openstackgerritAndreas Jaeger proposed openstack/releases master: Release pbr 1.11.0  https://review.openstack.org/43061807:52
*** AJaeger has joined #openstack-release07:58
AJaegerdhellmann, dims, can you help with this error, please? http://logs.openstack.org/18/430618/1/check/gate-releases-tox-validate-ubuntu-xenial/beb762a/console.html#_2017-02-08_07_54_11_32704107:58
* AJaeger does not know what's wrong with pbr here07:59
AJaeger"python setup.py --name" works fine for me in pbr08:00
*** amoralej|off is now known as amoralej08:04
*** sdake_ is now known as dake08:05
*** sdake has joined #openstack-release08:14
*** dake has quit IRC08:17
*** david-lyle_ has quit IRC08:28
*** udesale has joined #openstack-release08:30
*** stevemar has quit IRC08:33
*** stevemar has joined #openstack-release08:35
*** jpich has joined #openstack-release08:46
*** amotoki_ has quit IRC09:09
*** amoralej is now known as amoralej|brb09:23
*** ricolin has quit IRC09:24
*** openstackgerrit has quit IRC09:34
*** lucas-afk is now known as lucasagomes09:34
*** ricolin has joined #openstack-release09:38
*** abadger2000 has joined #openstack-release09:40
*** Guest66676 has joined #openstack-release09:41
*** lbragstad_ has joined #openstack-release09:43
*** dims_ has joined #openstack-release09:44
*** dansmith_ has joined #openstack-release09:44
*** coreycb` has joined #openstack-release09:44
*** askb has quit IRC09:45
*** dims has quit IRC09:45
*** jtomasek has quit IRC09:45
*** dansmith has quit IRC09:45
*** skraynev has quit IRC09:45
*** abadger1999 has quit IRC09:45
*** lbragstad has quit IRC09:45
*** coreycb has quit IRC09:45
*** Guest66666 has quit IRC09:45
*** dansmith_ is now known as dansmith09:45
*** kzaitsev_mb has joined #openstack-release09:46
*** dansmith is now known as Guest4449209:46
*** jtomasek has joined #openstack-release09:52
*** askb has joined #openstack-release09:52
*** skraynev has joined #openstack-release09:52
*** wolverineav has joined #openstack-release09:55
*** wolverineav has quit IRC09:59
*** wolverineav has joined #openstack-release10:05
*** wolverineav has quit IRC10:16
*** yanyanhu has quit IRC10:19
*** amoralej|brb is now known as amoralej10:20
*** udesale has quit IRC10:34
*** wolverineav has joined #openstack-release10:36
*** wolverineav has quit IRC10:41
*** claudiub has quit IRC10:42
*** wolverineav has joined #openstack-release10:53
*** wolverineav has quit IRC10:57
*** openstackgerrit has joined #openstack-release11:01
openstackgerritMerged openstack/releases master: Release Puppet OpenStack Ocata-rc1  https://review.openstack.org/43041511:01
*** electrofelix has joined #openstack-release11:03
*** sdake has quit IRC11:11
*** sdake has joined #openstack-release11:11
*** claudiub has joined #openstack-release11:12
*** wolverineav has joined #openstack-release11:15
*** wolverineav has quit IRC11:20
*** wolverineav has joined #openstack-release11:34
*** sdague has joined #openstack-release11:38
*** wolverineav has quit IRC11:39
*** sdake has quit IRC11:43
*** wolverineav has joined #openstack-release11:54
*** ricolin has quit IRC11:54
ianychoiHi release team! Is there a kind of tool such as 1) to see the list of projects which have stable/ocata branch and 2) list of projects for cycle-with-milestone, cycle-with-intermediary, cycle-trailing, and independent?11:55
*** wolverineav has quit IRC11:59
*** pbourke has quit IRC12:06
*** pbourke has joined #openstack-release12:07
*** dave-mccowan has joined #openstack-release12:08
*** wolverineav has joined #openstack-release12:14
*** mburned_out is now known as mburned12:17
*** wolverineav has quit IRC12:18
*** lucasagomes is now known as lucas-hungry12:22
dhellmannianychoi : yes, in the releases repo there is a list-deliverables command. Try "tox -e venv -- list-deliverables -h"12:22
dhellmanndims, ttx, mtreinish, tonyb, fungi : o/12:22
ianychoidhellmann, wow great thanks a lot :)12:22
ianychoidhellmann, now I need to more think about Soft & Hard StringFreeze periods and different release models.. so I want to see such list12:23
dhellmannAJaeger : could this be related? http://logs.openstack.org/18/430618/1/check/gate-releases-tox-validate-ubuntu-xenial/beb762a/console.html#_2017-02-08_07_54_11_28201612:23
*** wolverineav has joined #openstack-release12:24
*** david-lyle has quit IRC12:24
AJaegerdhellmann: could be ;(12:26
AJaegerharlowja: could you check the above pbr problem, one of your changes introduced that ^12:28
*** wolverineav has quit IRC12:28
dhellmannAJaeger : the validation job worked for me locally12:30
AJaegerdhellmann: so, recheck?12:31
dhellmannyeah, I just did12:31
dhellmannit's very likely that my system is different from the CI system though12:31
dhellmannI'm sure I have pbr installed globally, for instance12:31
AJaegerok12:31
AJaegersame problem again, dhellmann ;(12:35
AJaegerso, we need to file a bug with oslo team...12:35
AJaegerdhellmann: can you add details on how to reproduce to https://bugs.launchpad.net/pbr/+bug/1662877 , please?12:38
openstackLaunchpad bug 1662877 in PBR "pbr HEAD broken" [Undecided,New]12:38
AJaegerharlowja: ^12:38
dhellmannAJaeger : right now the only way to reproduce it is in that job :-)  I'll see if I can figure out what's going on12:38
AJaegerthanks, dhellmann.12:38
dhellmannAJaeger: fwiw, one of the reasons work on fixing that warnerrors flag stalled was that restoring it as-is will likely break a lot of doc builds. So we need to be careful about it. I thought of maybe adding a flag with a different name so we could release the fix and then roll it out across projects.12:40
dhellmannI don't know what the current fix looks like12:40
*** kzaitsev_mb has quit IRC12:42
AJaegeryeah, that is indeed a challenge...12:46
*** coreycb` has quit IRC12:48
*** coreycb has joined #openstack-release12:48
*** kzaitsev_mb has joined #openstack-release12:53
*** wolverineav has joined #openstack-release12:53
openstackgerritDoug Hellmann proposed openstack/releases master: Release pbr 1.11.0  https://review.openstack.org/43061812:53
openstackgerritDoug Hellmann proposed openstack/releases master: use tox env when asking pbr for its name  https://review.openstack.org/43088912:53
dhellmannAJaeger : let's see if ^^ helps, although if it does that won't fix the actual release jobs12:53
AJaegerinteresting...12:54
dhellmannit's a guess, based on an import error happening when asking pbr for its name12:55
dhellmannand the fact that it worked when I ran it locally12:55
AJaegerlet's see what happens ;)12:56
* dhellmann nods12:56
* AJaeger nods - it passes according to zuul status page ;)12:56
AJaegerdhellmann: thanks a lot12:57
AJaegerdhellmann: I leave the change as WIP for a few more days - do you have an idea on how to go ahead with the warnerrors. Projects should be able to fix their stuff quickly if needed or change warnerrors - still, not a nice thing12:58
*** wolverineav has quit IRC12:58
dhellmannAJaeger : ok, so then repro steps should be to run "python setup.py --name" in the pbr repository on a system without pbr already installed12:58
dhellmannyeah, the best I could come up with is changing the name of the flag12:58
*** askb has quit IRC12:58
AJaegerdo you want to start a discussion - or just do it?12:59
dhellmannAJaeger : let's just change the name, and then when it's released we can have the discussion about rolling out the change13:01
dhellmannI don't know if having something in setup.cfg that pbr doesn't know about will be an error, so we may need to add a flag and just ignore the old one13:01
dhellmannI don't know how that works off the top of my head13:01
AJaegerdhellmann: will you have time for that change? Sorry, I can't this week ;(13:02
dhellmannAJaeger : it has been on my list for a while now, but probably not this week either13:02
dhellmannI've been meaning to get to it for the whole cycle13:03
AJaegerdhellmann: ok13:03
AJaegeroops ;(13:03
*** bobh has joined #openstack-release13:08
*** bobh has quit IRC13:09
dhellmannsdague: are we ready to branch devstack?13:09
sdaguedhellmann: go for it13:11
dhellmannok, doing that now13:11
dhellmanngrenade is up next13:11
dhellmannrunning: ./make_branch.sh openstack-dev/devstack stable/ocata HEAD13:12
*** wolverineav has joined #openstack-release13:13
dhellmannoops, have to install zuul-cloner locally first13:13
dhellmannthere we go13:13
dhellmannand the .gitreview update for devstack is in https://review.openstack.org/43089713:13
dhellmannfull job output is http://paste.openstack.org/show/598083/13:14
dhellmannsdague : I don't see anything in the gate for grenade, are there any pending reviews we need to wait for?13:15
AJaegerdhellmann: the proposal node is currently offlined a few times...13:16
sdaguedhellmann: no, grenade should be fine13:16
AJaegerwe first need to get all the current periodic jobs finished, then it should be fine again13:16
dhellmannAJaeger : I'm not sure what that means?13:16
AJaegerdhellmann: I think one of your jobs runs on the proposal node - and that node is currently busy with translations13:16
dhellmannAJaeger : normally that's true. For some of these repos I have to run the commands locally by hand because we don't tag them13:17
AJaegerAnd since we had a bug in the new ocata translation jobs, those will "offline" the proposal node afater run - and we need to start it again13:17
dhellmannoh, so the jobs are taking the node offline13:17
*** wolverineav has quit IRC13:17
dhellmannbranching grenade with: ./make_branch.sh openstack-dev/grenade stable/ocata HEAD13:18
dhellmann.gitreview update is in https://review.openstack.org/43090513:18
dhellmannjob output is in http://paste.openstack.org/show/59808513:18
*** lucas-hungry is now known as lucasagomes13:18
*** trown|outtypewww is now known as trown13:19
dhellmannthe next step is to land the devstack-gate changes in https://review.openstack.org/#/c/427916/113:19
dhellmannI've rechecked that patch, since it failed before (maybe because the branches didn't exist?)13:19
dhellmannoops, sorry, that's the grenade settings patch not d-g13:19
dhellmannbut it is next13:20
*** tommylikehu_ has joined #openstack-release13:26
*** wolverineav has joined #openstack-release13:26
*** trown has left #openstack-release13:27
* dhellmann steps away to make breakfast while those tests run13:28
*** lbragstad_ is now known as lbragstad13:29
*** wolverineav has quit IRC13:31
*** amoralej is now known as amoralej|lunch13:34
*** ricolin has joined #openstack-release13:36
*** sdake has joined #openstack-release13:36
*** yuvalb has joined #openstack-release13:42
openstackgerritYuval Brik proposed openstack/releases master: Release karbor-dashboard 0.1.2  https://review.openstack.org/43092313:43
openstackgerritYuval Brik proposed openstack/releases master: Release python-karborclient 0.1.3  https://review.openstack.org/43092413:43
openstackgerritMerged openstack/releases master: Add a whitespace between words  https://review.openstack.org/43052413:46
ttx dhellmann o/13:51
dhellmannttx: o/13:52
dhellmannsdague : I'm likely going to need some help debugging the job failures on https://review.openstack.org/42791613:52
sdaguedhellmann: well the branches don't exist yet right?13:53
sdagueWe need those branches, then we can recheck13:53
sdagueor are they failing now?13:53
dhellmannwhich branches? we should have branches for devstack and grenade13:53
dhellmannoh, they're failing again right now13:53
dhellmannI'm looking at zuul's ui, it hasn't reported back to gerrit13:53
*** dtantsur has joined #openstack-release13:55
*** gordc has joined #openstack-release13:55
dhellmannI wonder if maybe some of the steps are out of order13:56
*** zhurong has joined #openstack-release13:57
dhellmannsdague : looks like some sort of policy or authorization issue? http://logs.openstack.org/16/427916/1/check/gate-grenade-dsvm-neutron-ubuntu-xenial/56b7a7d/logs/grenade.sh.txt.gz#_2017-02-08_13_48_57_24914:00
sdaguedhellmann: http://logs.openstack.org/16/427916/1/check/gate-grenade-dsvm-neutron-ubuntu-xenial/56b7a7d/logs/apache/keystone.txt.gz#_2017-02-08_13_48_57_24714:04
sdagueI'm kind of confused14:04
openstackgerritAlex Schultz proposed openstack/releases master: Release puppet-ceph 2.3.0  https://review.openstack.org/43093714:04
sdaguedtroyer: any idea why http://logs.openstack.org/16/427916/1/check/gate-grenade-dsvm-neutron-ubuntu-xenial/56b7a7d/logs/grenade.sh.txt.gz#_2017-02-08_13_48_54_600 doesn't seem to be working14:05
sdaguedhellmann: it looks like openstack client return fields changed?14:09
dhellmannugh14:09
sdaguedhellmann: http://logs.openstack.org/16/427916/1/check/gate-grenade-dsvm-neutron-ubuntu-xenial/56b7a7d/logs/grenade.sh.txt.gz#_2017-02-08_13_48_54_60014:09
sdaguethe id field for that create isn't the user_id14:09
sdaguewhich is what it used to be14:09
dhellmannstevemar : any ideas? ^^14:10
stevemareh14:10
sdagueI actually have no idea what that id is14:11
sdaguehttp://logs.openstack.org/16/427916/1/check/gate-grenade-dsvm-neutron-ubuntu-xenial/56b7a7d/logs/apache/keystone.txt.gz#_2017-02-08_13_48_55_962 is where it's created14:11
sdaguee8... is the userid created14:11
sdagueI have no idea where the d2... is coming from14:11
*** amoralej|lunch is now known as amoralej14:13
*** mriedem has joined #openstack-release14:13
sdagueit isn't findable anywhere else in the logs14:13
stevemarthe user create command shouldn't have changed14:13
stevemari see u'e80d2c25ca66484d809fa06efbef999d14:13
sdaguestevemar: yeh14:13
sdaguebut openstack client doesn't return that14:14
sdaguestevemar: we are running - openstack user create nova_grenade --project d435b3cefa514338a2451afe39fc54c0 --password pass -f shell -c id14:14
sdagueand that call is returning a uuid I can't find anywhere else14:14
sdagueand definitely not e80..14:14
stevemarsdague: that e80 may be the user that initiated the call14:15
sdaguestevemar: um what now?14:15
stevemarsdague: https://github.com/openstack/keystone/blob/master/keystone/policy/backends/rules.py#L73-L7814:15
stevemarsdague: not the user that was created, but the user that did the creating14:16
stevemarthats why you see u'e80d2c25ca66484d809fa06efbef999d all over the place14:16
sdaguestevemar: ah ok14:16
sdagueso what that says to me is the user create failed14:16
sdaguebut says it succeeded14:17
stevemarit looks like the OSC command passed to me, http://logs.openstack.org/16/427916/1/check/gate-grenade-dsvm-neutron-ubuntu-xenial/56b7a7d/logs/grenade.sh.txt.gz#_2017-02-08_13_48_56_034 says it created it with id = d221effd89844fbf9b959892e975ac2f14:17
sdagueopenstack user create nova_grenade --project d435b3cefa514338a2451afe39fc54c0 --password pass -f shell -c i14:17
sdaguestevemar: right14:17
sdagueand then keystone fails 2 seconds later14:17
sdagueand sasy d22.. has no access to project d43...14:17
sdaguehttp://logs.openstack.org/16/427916/1/check/gate-grenade-dsvm-neutron-ubuntu-xenial/56b7a7d/logs/apache/keystone.txt.gz#_2017-02-08_13_48_57_24614:18
sdagueso... something failed14:18
stevemarah and d435b3cefa514338a2451afe39fc54c0 was passed in as the 'default project'14:18
sdaguestevemar: so, we've seen this pattern before, the is the first time that openstack client / ocata is being used in grenade for building these things14:19
sdagueso the question, the previous command to accomplish this is: openstack user create nova_grenade --project d435b3cefa514338a2451afe39fc54c0 --password pass -f shell -c id14:20
sdagueis that command no longer valid?14:20
dtroyerI dion't recall anything around user/project CRUD changing recently14:21
sdagueso, in newton, that command does both create and role assignment14:22
sdaguedevstack does role assignment in a separate step, but grenade never did14:22
stevemarsdague: i think it may be an issue with the v3 switch over, in v3 the "default role assignment" bit isn't as friendly, we prefer it being explicitly set14:23
sdaguestevemar: ok, which is definitely a user visible regression14:23
dhellmannwouldn't that change have affected going from newton to ocata? how is this a new failure?14:23
*** udesale has joined #openstack-release14:23
sdaguedhellmann: because we don't create new users on the "new" side14:24
dhellmannah14:24
sdagueat least not with this invocation14:24
dhellmannmakes sense14:24
sdagueok, I can put a work around into grenade14:24
sdaguebut... this definitely feels like a user visible regression14:25
dhellmannsdague : thanks14:25
dhellmannis that a change in osc or in keystone?14:25
sdagueI honestly don't know14:25
dhellmannI'm guessing osc?14:25
stevemardhellmann: checking the server side code now14:25
dhellmannstevemar : thanks14:25
*** udesale has quit IRC14:27
*** udesale has joined #openstack-release14:27
*** Guest44492 is now known as dansmith14:28
stevemardhellmann: sdague oh you're going to love this14:28
dhellmannsomething tells me I'm not :-)14:28
*** dansmith is now known as Guest9404614:28
sdagueso, actually, I'm now completely confused by the role call I need to make here, if dtroyer or stevemar could do the grenade patch that might be best14:30
dims_ttx : dhellmann : o/ sorry late start today14:30
dhellmanndims_ : ack, np, I wanted to try to get a head start before the gate got backed up14:31
dhellmannI'm going to have to step away for a few min in about 15-20 min14:31
* dims_ reading backscroll14:31
dtroyerI'm heading back into all-day f2f meetings… but this isn't a change in OSC that I can think of.  when stevemar talks about a v3 default domain change, that's a Keystone v3 thing that v2 used to do that v3 doesn't14:31
stevemardhellmann: sdague i'm not seeing that we ever do anything with the 'default_project_id' in the v3 create_user manager, aside from validate it exists: https://github.com/openstack/keystone/blob/master/keystone/identity/core.py#L918-L91914:31
dhellmanndtroyer : ack, thanks14:32
dhellmannstevemar, sdague : would it make sense to tell osc to use v2 then?14:32
stevemardhellmann: i would just add the necessary call14:32
sdaguestevemar: ok, so basically all the explicitness changes in devstack covered up this behavior change?14:32
dhellmannstevemar : to what? grenade?14:32
sdaguestevemar: what is that call going to be?14:32
stevemarsdague: yes14:33
stevemardhellmann: yes, to grenade14:33
stevemarsdague: one moment, just looking up a reference example for ease14:33
sdaguestevemar: because all the devstack osc logic is about creating new roles in the abstract, and honestly my brain is too crufty on all this to make sense of it14:33
dhellmannstevemar: I can file a bug for this against keystone while you work with sdague on that grenade patch14:34
stevemardhellmann: to add logic to keystone to add a default role?14:34
* dtroyer sees a market for a "get me a project/user" business-logic function for osc/shade/oaktree14:35
dhellmannstevemar : we need to make grenade set up the environment properly, so if we can add an osc call there to do that I think that unblocks us the quickest14:35
*** Guest94046 is now known as dansmith14:35
* dhellmann thinks our apis are being too pedantic about separation of responsibilities14:35
dtroyer(from memory) I think we did have to do this all to devstack, long ago, to make it work properly with v314:35
sdaguedtroyer: yeh, there is a ton of complexity there14:36
sdaguedtroyer: it just does seem sensible that the API for a user create would involve default project14:37
sdaguedhellmann: yeh, 100% agree14:37
dhellmannstevemar : the keystone bug for later is https://bugs.launchpad.net/keystone/+bug/166291114:38
openstackLaunchpad bug 1662911 in OpenStack Identity (keystone) "v3 API create_user does not use default_project_id" [Undecided,New]14:38
dtroyerFor context, this is part of branching devstack & grenade, and is in master?14:41
dhellmanndtroyer : yes14:41
dhellmannhttps://review.openstack.org/#/c/427916 is the patch with jobs failing14:42
stevemarsdague: should just need to run `openstack role add member --user fcc0b75ff7954f81b2aa713b0a4cd972 --project 11743ad9e5884f28bc4bc883d7005104`14:42
dhellmannthough the log links on that page are old, it's still in the check queue14:42
sdaguestevemar: ok, there is automatically a role for the project id?14:42
stevemarsdague: thats "member"14:42
dhellmannthe role named "member" doesn't have to be created separately?14:43
stevemarshold i modify https://github.com/openstack-dev/grenade/tree/master/projects/10_keystone ? and add from-ocata14:43
sdaguewait, so all users end up in member?14:43
sdaguethe error is that $user doesn't have permissions for $project14:44
stevemarsdague: dhellmann over leveraging the "member" role was part of the issue we moved away from giving users a default project14:44
mtreinishdhellmann: sorry, just got online. Do you need help with the branching stuff?14:44
sdagueI want to make sure we aren't papering over this in a different way that is so broad we're letting some other thing fall through14:44
dhellmannmtreinish : np, we're currently debugging an issue with keystone/osc/grenade14:45
sdaguestevemar: so that means that all added users / projects have permersions against each other?14:45
stevemarsdague: the user doesn't any permission anywhere14:45
sdaguestevemar: right. Ok, here is the hypothetical14:45
sdaguecreate project114:46
sdaguecreate project214:46
sdaguecreate user114:46
sdaguecreate user214:46
sdaguerole add member user1 project114:46
sdaguerole add member user2 project214:46
sdaguecan I do actions as user1 on project2 resources, if I pass project2 id?14:46
stevemarshouldn't be able to get a token14:47
sdagueok, so the roles and what tokens I can get are different logic paths?14:48
stevemaryep, when you ask for a token on a specific project it checks that the user has at least one role on it14:49
dhellmannah, so it doesn't matter which role as long as there is one14:49
sdaguestevemar: ok, so role add member user1 project114:50
stevemarbingo14:50
sdagueis adding member role on project1 for user1?14:50
stevemarsdague: correct14:50
sdaguenot member role being a top level thing, and project1 / user1 in it14:50
sdagueok14:50
stevemarright right, definitely not that14:50
dhellmannit feels like user and project should be positional args for that call14:51
dhellmannI guess since they're guids it's a little more friendly to spell out which is which14:51
stevemardhellmann: i believe they are xors? --user|--group and --project|--domain14:51
dtroyerFWIW, this was also the most confusing thing we did in either of the UX studies… operators still struggle with it.14:51
dhellmannstevemar : oh, ok, I didn't realize there were other options14:51
dtroyerdhellmann:  its a 3-was association and I think we landed on role being the primary resource14:52
*** kgiusti has joined #openstack-release14:52
dhellmanndtroyer : I wonder if separate commands would make sense, with "project role" and "domain role" being different resources? I don't know the data model well enough, so I should probably stop redesigning the UI14:52
openstackgerritYuval Brik proposed openstack/releases master: Release Karbor 0.1.1  https://review.openstack.org/43097714:53
dtroyerqualifying role would be an artificial thing for keystone but might be a cognitive help to users, I hadn' thought of it that way14:53
dhellmanndtroyer : right, that would be just for the users' benefit14:54
sdagueit might be, but it would also be good to just add user to a project by default14:54
dhellmannsdague : that, too14:54
sdagueI mean, in unix you can't get a user without a default group set14:54
sdagueso it's massively confusing to find out that you can in openstack users14:54
dhellmannI think that's the keystone v3 bug, though, isn't it?14:54
dtroyersdague: agreed… even the idea of a default domain was pushed against to begin with14:54
sdaguedhellmann: yeh14:54
dtroyerso if clients do that, we have to be able to configure the default14:54
dhellmannso we've got that filed, and we know what to do in grenade to move to our next step14:55
sdaguedhellmann: yeh, I'm writing the patch now14:55
dhellmannsdague : yeah, just making sure I know what's going on :-)14:55
dims_:)14:55
dtroyerI have to make a venue change in a minute…14:55
dhellmanndtroyer : thanks, we'll chain stevemar to his desk for the short-term14:55
dhellmannoh, speaking of which, I need to step away for 5-10 min. bbiab14:56
* dtroyer forgot where he stashed his copy of that key14:56
stevemari'm actually heading out to a hospital in about an hour :(14:56
dtroyerI'll be back by then so you can tag me14:56
stevemari will gladly substitute myself with an lbragstad :P14:56
dtroyeroh!  that's official!  congrats? lbragstad :)14:58
stevemardtroyer: it is!14:59
stevemardhellmann: going back to the history of it, part of the reasoning for not creating a role by default is because we didn't want to have to create the "member" role by default, since it was yet another hard coded thing we had in v214:59
stevemarwhich led to hardcoded things in policy files15:00
stevemarwhich continues to be a huge PITA for us to make any headway on15:00
sdaguestevemar: well, I think you just hotswapped PITAs :)15:00
dtroyerstevemar: is there a default_role in keystone.conf like domain?15:00
stevemarsdague: yup :)15:01
dtroyerif the default swere available via the API, a "smart" client could do that for you15:01
sdaguestevemar: I would actually say that's not really the major issue with moving forward on policy, but that's for another time15:01
dtroyer"get me a user"15:01
stevemardtroyer: you're thinking of these, which we state are v2 only: https://github.com/openstack/keystone/blob/master/keystone/conf/default.py#L92-L11215:02
sdague https://review.openstack.org/430982 I think is the fix15:02
sdaguedtroyer: ..... *sigh*15:02
stevemardtroyer: we could open it up15:02
stevemarsdague: +115:04
*** adrian_otto1 has joined #openstack-release15:05
* dhellmann returns15:07
dhellmannif the default is in keystone.conf, why should the client have to ask for it? why not just add the role?15:09
mwhahahadhellmann: quesiton on http://logs.openstack.org/37/430937/1/check/gate-releases-tox-validate-ubuntu-xenial/b58d94e/console.html#_2017-02-08_14_17_28_785940 as puppet-ceph doesn't follow the stable/<openstack cycle> branching, do i need to update some metadata somewhere to skip that check?15:15
sdaguedhellmann: I updated the d-g patch with a depends-on, that should prove it one way or another - https://review.openstack.org/#/c/427916/15:17
*** dave-mccowan has quit IRC15:20
*** adrian_otto1 is now known as adrian_otto15:21
dims_thanks sdague15:22
dhellmannmwhahaha : is there no newton branch, then?15:23
mwhahahadhellmann: no, it follows the ceph release so stable/hammer is the last release15:23
dhellmannsdague : thanks15:23
dhellmannmwhahaha : ok. we'll have to put in an exception or something15:24
dhellmannmwhahaha : is it switching to using our series names?15:24
mwhahahadhellmann: no15:25
dhellmannmwhahaha : ok. none of the validation for creating the branch is going to work either, then. you may want to just manage that one by hand.15:26
*** hongbin has joined #openstack-release15:26
mwhahahadhellmann: ugh15:26
dhellmannmwhahaha : nature abhors a vacuum, and I abhor a special case15:27
sdagueapparently, wrong osc syntax15:28
sdaguenew patch15:28
mwhahahadhellmann: i would argue that the release process is not flexible enough to handle auxiliary components. this was here previously but what was added was more assumptions that made things mandatory that weren't prviously15:28
mwhahahadhellmann: it seems like the branching naming strategy should be configurable (ie can be disabled)15:28
dhellmannmwhahaha : we have 3 people on the release team. we're using automation to handle the number of projects. the more consistent those projects are, the easier it is for everyone to learn and understand the process. you're not prevented from having a special case, but we're trying to make you take on that load instead of us.15:29
mwhahaha3 people > 115:30
mwhahahabut ok15:30
dhellmannyou're 1 person with 1 special case15:30
mwhahahathat existed before you wrote the new requirements15:30
dhellmannthis is no different from when we didn't handle the releases for you at all15:30
dhellmannthe fact that it's an old case doesn't make it less unique15:30
mwhahahahow do the oslo libs get handled hten15:31
mwhahahado they still have stable/newton?15:31
mwhahahaetc?15:31
dhellmannyes15:31
*** mriedem1 has joined #openstack-release15:31
dhellmannwe added the metadata for the branch locations for projects that had stable/newton branches so this check would pass15:31
dhellmannthe new check is to help avoid errors as we move into pike. we've already had 2 cases where releases happened on the wrong branch15:32
mwhahahafine i'll deal with it myself then15:33
dhellmannmwhahaha : it might work if you move that deliverable to the _independent directory, I think the check is disabled there15:33
dhellmannI think that's how we deal with gnocchi, but it was already independent15:33
*** mriedem has quit IRC15:34
dtroyerdhellmann: in this case right, a "use default" would suffice, but there are other places having that info would be very helpful in hiding the ugliness15:35
* dtroyer is back too15:35
dhellmanndtroyer : if this is a named thing, maybe it should just be hard-coded in the same way that we standardize the names of service catalog entries15:35
sdaguedtroyer / stevemar you want to validate I'm actually using the osc command correctly? Because I failed the first time :)15:36
dhellmannit's not quite the same case, it just seems like making it flexible is going to cause interop issues unless we also force clients to be smarter about something15:36
sdaguehttps://review.openstack.org/#/c/430982/15:36
dtroyersdague: looking now15:36
*** wolverineav has joined #openstack-release15:37
dtroyerdhellmann: exactly, and this argument is what was had a couple of years ago when those decisions were made.  (I wasn't there either)15:37
dhellmannit still makes me sad that we've had to build a "usable API service" on top of the regular ones15:37
sdagueyeh, I think there was a pendulum swing towards flexibility before there were enough users to see how much it sucked for them :)15:37
dtroyermee too dhellmann15:37
dtroyerI'm pushing for more opinions not less15:38
*** dave-mccowan has joined #openstack-release15:38
*** mriedem1 is now known as mriedem15:38
sdaguethat swing is coming back, but the 2 to 3 cycle (or more) delay on people rolling out updates means there is a ton of innertia there15:38
dtroyersdague: that looks good at a glance, I'm going to go try it though15:39
sdaguedtroyer: coolio15:39
sdagueit takes about 25 minutes to get to that chunk of code15:39
*** wolverineav has quit IRC15:41
stevemarsdague: now we gotta fix the whole "admin" user, project, role, and policy check being hard-coded too :)15:44
sdaguestevemar: so, while I'm fine with scoping there, every time I've been in a room with big users they pretty much aren't concerned about it15:48
sdagueway more concerned about getting an observer role baked in15:49
sdaguefor monitoring tools15:49
openstackgerritMerged openstack/releases master: Release karbor-dashboard 0.1.2  https://review.openstack.org/43092315:52
openstackgerritMerged openstack/releases master: Release python-karborclient 0.1.3  https://review.openstack.org/43092415:53
* dhellmann has to step away again15:53
*** wolverineav has joined #openstack-release15:57
*** inc0 has joined #openstack-release15:57
*** adrian_otto has quit IRC15:57
*** wolverineav has quit IRC16:01
dims_sdague : 7 failures? (http://logs.openstack.org/16/427916/3/check/gate-grenade-dsvm-neutron-multinode-ubuntu-xenial/78f0558/logs/grenade.sh.txt.gz)16:02
*** armax has joined #openstack-release16:03
dims_"is not mapped to any cell"16:03
sdaguedims_: ok, that's different though16:04
sdaguedtroyer: gah, role member does not exist.... :(16:04
*** Rockyg has joined #openstack-release16:04
* dtroyer goes to look16:06
sdaguehttp://logs.openstack.org/82/430982/2/check/gate-grenade-dsvm-neutron-ubuntu-xenial/42e0eaa/logs/grenade.sh.txt.gz#_2017-02-08_15_49_47_87016:07
dtroyerso WTF?  DevStack creates Member, but "os role add member --user admin --project demo" didn't throw chunks16:07
*** kzaitsev_mb has quit IRC16:08
sdagueit's "Member" .... o_O16:09
openstackgerritWitold Bedyk proposed openstack/releases master: Release monasca-agent 1.7.0  https://review.openstack.org/43103216:09
dtroyerright.  just like domain.  but "member" doesn't throw an error for me…16:09
dtroyeror Domain16:10
sdagueit threw an error here16:10
sdaguehonestly, who knows, new patch pushed16:10
dtroyerthat's why I wanted to run it myself… now I'm doubting my ability to have a reasonably representative DevStack16:10
sdagueyeh, so in devstack we only reference these things by id16:13
sdaguewhich, I think is another time where this whole lots of osc wrappers in devstack covered up an incompatibility16:13
*** kzaitsev_mb has joined #openstack-release16:15
*** adrian_otto has joined #openstack-release16:16
dhellmannI guess the role id isn't available to use instead of the name?16:19
dhellmannprobably not if devstack created the role16:20
sdaguedhellmann: we could make a dedicated call to get that16:20
sdaguehowever... it actually seems massively not the way anyone would ever use these tools16:20
dhellmannright16:20
dhellmannand if that call needs the name anyway, it doesn't help16:21
sdagueoh, it can use the id16:21
*** david-lyle has joined #openstack-release16:23
*** udesale has quit IRC16:24
*** zhurong has quit IRC16:25
sdaguemeh, I'm going to make a devstack change to use strings and see how all this fails16:29
*** sdake_ has joined #openstack-release16:36
*** sdake has quit IRC16:38
*** kzaitsev_mb has quit IRC16:41
*** lennyb has quit IRC16:44
*** ricolin has quit IRC16:50
*** tommylikehu_ has quit IRC16:51
*** tommylikehu_ has joined #openstack-release16:52
*** tommylikehu_ has quit IRC16:52
*** david-lyle has quit IRC16:52
*** tommylikehu_ has joined #openstack-release16:52
*** Guest66676 is now known as Guest666616:53
*** tommylikehu_ has quit IRC16:53
*** tommylikehu_ has joined #openstack-release16:53
*** tommylikehu_ has quit IRC16:54
*** tommylikehu_ has joined #openstack-release16:54
*** tommylikehu_ has quit IRC16:55
*** tommylikehu_ has joined #openstack-release16:55
*** tommylikehu_ has quit IRC16:55
*** david-lyle has joined #openstack-release17:12
*** mriedem has quit IRC17:12
*** mriedem has joined #openstack-release17:14
*** armax has quit IRC17:23
*** armax has joined #openstack-release17:24
ttxdhellmann: just processed the last one that seemed ready, will log off soon17:31
*** armax has quit IRC17:32
sdaguethe d-g change + grenade tests have just gotten to the critical part. If it passes I'm going to approve the grenade change and go for a walk (it's 50+ outside today, and we're going to get 6 - 12 inches of snow tomorrow)17:34
sdagueoh, it looks like it got past17:34
openstackgerritMerged openstack/releases master: Release monasca-agent 1.7.0  https://review.openstack.org/43103217:35
sdaguedhellmann: just find another +2 for - https://review.openstack.org/#/c/42791617:35
dhellmannsdague : thanks17:37
*** lucasagomes is now known as lucas-afk17:37
*** dtantsur is now known as dtantsur|afk17:37
dhellmanndtroyer, mtreinish : you both have +2 on devstack-gate, right? https://review.openstack.org/#/c/42791617:38
dhellmannttx: ack, thanks, we'll keep going17:38
mtreinishdhellmann: I do not, nor does dtroyer iirc17:38
mtreinishdhellmann: it's infra-core and sdague iirc17:39
dhellmannoh, ok17:39
*** adrian_otto1 has joined #openstack-release17:39
ttxdhellmann: in other news looks like we are good to go on https://review.openstack.org/42975317:39
*** armax has joined #openstack-release17:39
dhellmannttx: I guess we want to do that before branching requirements17:40
ttxyeah, that's what i was thinking17:40
ttxexcept tonyb's -2 might prevent us17:40
ttxgot to run17:40
dhellmannyeah, it will17:40
dhellmannso that'll have to wait for him to come online later17:41
ttxor redo it17:41
dhellmannor that17:41
*** adrian_otto has quit IRC17:42
dtroyerdhellmann: ya, no, I'm not +2 on devstack-gate17:45
*** wolverineav has joined #openstack-release17:47
dhellmanndtroyer : jeblair got it17:49
sdaguedims_: ah, yeh, your cell issue is still there on multinode17:50
sdaguemriedem / dansmith - http://logs.openstack.org/16/427916/3/check/gate-grenade-dsvm-neutron-multinode-ubuntu-xenial/2ef2ed1/logs/grenade.sh.txt.gz#_2017-02-08_17_35_44_082 ?17:50
*** adrian_otto1 has quit IRC17:51
*** wolverineav has quit IRC17:52
dansmithsdague: context?17:52
sdaguechange that is trying to fix the test branches17:52
sdaguefor stable/ocata17:52
sdaguestable/ocata -> master fails to run multinode on grenade17:53
*** jpich has quit IRC17:53
*** adrian_otto has joined #openstack-release17:53
sdagueDetails: {u'message': u"Host 'ubuntu-xenial-2-node-rax-ord-7185115-417164' is not mapped to any cell", u'code': 400, u'created': u'2017-02-08T17:35:37Z'}17:53
sdaguein base tempest run17:53
dansmithif stable/ocata is the base, then we should be getting normal devstack setup of cells and things right?17:54
mriedemso tha's grenade starting with stable/ocata nova compute, going to nova master compute right?17:54
sdagueyes17:55
dansmithlike, why would base grenade not pass and devstack on ocata would?17:55
mriedemthere is no from_ocata in grenade or anything17:55
dansmithshouldn't need one because of base devstack no?17:55
sdaguedansmith: because where orchestration happens is different in the 217:55
mriedemyeah devstack on the base (ocata) side should do the things17:55
sdaguemriedem: https://github.com/openstack-infra/devstack-gate/blob/c435a724bd257b22a2e39e8e9125c11302a8c81d/devstack-vm-gate.sh#L71317:56
sdagueit's that missing call right?17:56
mriedemthat happens in devstack17:57
dansmithis that the thing we tried to land for both and then dropped to only devstack?17:57
mriedemvia simple_cell_setup17:57
dansmithmriedem: not after the subnode in grenade17:57
dansmithwe run it, but it might be too early17:58
sdaguemriedem: right, I'm trying to find the review that had those comments in it17:58
mriedemhttps://github.com/openstack-dev/devstack/blob/master/lib/nova#L952 is where devstack would discover it17:59
mriedembut,17:59
* dhellmann wonders if we should have a master->master grenade job to catch these sorts of things17:59
mriedemhttps://github.com/openstack-dev/devstack/blob/master/stack.sh#L137117:59
mriedemthat's not called if n-api isn't running17:59
mriedemso yeah, we don't discover the host on the subnode17:59
mriedemi could have sworn i was doing this at some point17:59
mriedemand got kicked in the nuts17:59
dansmithmriedem: you mean "we don't discover the subnode" right?17:59
dhellmannI have to drop offline to pick up a prescription. I'll be back in a bit.18:00
dansmithmriedem: you did, and we were asked to remove it fo the grenade side18:00
mriedemdansmith: n-cpu on the subnode right18:00
*** tommylikehu_ has joined #openstack-release18:00
sdaguemriedem: because there is no db on the subnode18:00
dansmithmriedem: you don't mean "run discover on the subnode" though18:00
mriedemhttps://review.openstack.org/#/c/420981/18:00
sdagueok, so this either needs to be here - https://github.com/openstack-infra/devstack-gate/blob/c435a724bd257b22a2e39e8e9125c11302a8c81d/devstack-vm-gate.sh#L65118:00
sdagueor directly in grenade after post-stack gets calls18:01
dansmithsdague: mriedem: https://review.openstack.org/#/c/424252/1/devstack-vm-gate.sh18:01
sdagueI remember having that conversation, I didn't realize that stuff didn't land18:01
dansmithline 66618:01
dansmith(ha)18:01
dansmithwe were asked to remove it for the grenade side18:01
dansmithbecause grenade knows about things,18:02
dansmithbut if we don't discover on base, we'll fail before we get to the upgrade in ocata18:02
dansmithnot in newton18:02
sdaguewell, actually that comment is exactly what I said here, it either needs to be in the post-stack.sh18:02
sdagueor in grenade itself18:02
sdaguebecause we're not getting to the 666 line, the fail is earlier18:03
sdague663 fails, as smoke tests are run directly by grenade18:04
mriedemthe command for discover_hosts is pretty gross https://github.com/openstack-infra/devstack-gate/blob/c435a724bd257b22a2e39e8e9125c11302a8c81d/devstack-vm-gate.sh#L23818:04
mriedemto cram that in here i mean https://github.com/openstack-infra/devstack-gate/blob/c435a724bd257b22a2e39e8e9125c11302a8c81d/devstack-vm-gate.sh#L65118:05
mriedemi can...18:05
sdaguemriedem: well, it doesn't need to be ansiblized18:05
sdagueit's running in the grenade context18:05
mriedemok,18:05
*** tommylikehu_ has quit IRC18:05
mriedemlet me put something up18:05
dansmithI'm doing it already18:05
mriedemoy18:06
dansmithdo I need to do the stdbuf stuff?18:06
sdaguedansmith: no18:07
dansmithand sudo?18:07
sdagueno, I think you can just get away with18:08
sdague$BASE_RELEASE_DIR/tools/discover_hosts.sh18:08
sdagueactually BASE_DEVSTACK_DIR18:08
dansmithoh and it does its own existence check for manage now right?18:08
sdaguedansmith: that I don't know18:08
dansmithnot $BASE/old/devstack/tools/discover_hosts-.sh?18:08
dansmithit does, so that's good18:09
sdagueBASE_DEVSTACK_DIR is more generic18:09
sdaguebecause old / new isn't actually guarunteed, definitely not used during local runs18:09
sdaguealthough, I guess it's written from d-g18:09
sdagueso... yeh, $BASE/old/devstack/tools/discover_hosts-.sh? probably fine18:10
dansmithis there a bug?18:10
sdagueno18:10
sdaguewe've just been live debugging the transition issues since earlier this morning18:10
dansmithhttps://review.openstack.org/43109518:10
*** wolverineav has joined #openstack-release18:11
sdaguecool, rebased the branch change on top of it18:12
sdagueresults in an hour... walk time18:12
mriedemno18:12
mriedemyou can't just call discover_hosts.sh18:12
mriedembecause of big tent fun18:12
mriedemyou need something like "test -f tools/discover_hosts.sh" in there18:12
sdagueyou probably can here18:13
sdagueno one in the big tent uses multinode grenade that I know of18:13
mriedem(test -f tools/discover_hosts.sh && ./tools/discover_hosts.sh) || (! test -f tools/discover_hosts.sh)18:13
mriedemis essentially what d-g is doing18:13
dansmiththis is directly expecting devstack no?18:14
mriedemoh sorry,18:14
mriedemthe big tent thing was that discover_hosts was calling nova-manage when nova-manage wasn't installed,18:14
mriedemthe test check is for stable branches18:14
dansmithI'm not even sure I remember why we need that test in the other part18:14
dansmithah right18:14
mriedembecause discover_hosts.sh in devstack only existed in ocata18:14
mriedemit was to avoid a bunch of branch logic in d-g18:14
dansmiththe big tent check was the thing in discoer_hosts.sh itself18:14
mriedemyup18:14
mriedemi mean, it's probably a good acid test for which projects to drop, but you know... :)18:15
dansmithwell, ceilo failed that18:15
dansmithbut no argument from me :P18:15
*** wolverineav has quit IRC18:16
*** wolverineav has joined #openstack-release18:18
mriedemthey also failed pg18:20
*** wolverineav has quit IRC18:25
*** wolverineav has joined #openstack-release18:25
*** sripriya has quit IRC18:35
*** mwhahaha has quit IRC18:35
*** robcresswell has quit IRC18:35
*** kozhukalov has quit IRC18:35
*** igorbelikov has quit IRC18:35
*** dasanind has quit IRC18:35
*** nikhil has quit IRC18:35
*** kong has quit IRC18:37
dansmith2017-02-08 18:33:53.143 | ./post-stack.sh: line 7: /opt/stack/old/devstack/tools/discover_hosts.sh: No such file or directory18:40
dansmithis that not right? seems right18:41
* dhellmann returns 18:42
dansmithis that running against newton maybe?18:42
dansmithmust be I guess since the patch above it is the branch18:43
*** ashtokolov has quit IRC18:45
*** betherly has quit IRC18:45
*** sridhar_ram has quit IRC18:45
*** johnsom has quit IRC18:45
dhellmanndansmith : yeah, I think you're right, we're still trying to get d-g to understand about the ocata->master branch so patches are running newton->ocata for now18:45
dhellmannor maybe newton->master, which should be close to the same thing18:46
dansmithdhellmann: ack, I just made it only run it if it's there18:46
dhellmann++18:46
dhellmannI'd like to figure out how to predict this sort of stuff so the end of release process is a little smoother18:46
dhellmannwould a master->master grenade job, maybe just periodic, be useful for that?18:47
dansmithdhellmann: well, we actually kinda did catch this,18:47
dhellmannwell, by "catch" I really mean "avoid" or "deal with before the deadline"18:47
dansmithbut we dropped this call instead of moving, at least in part because I didn't understand some review feedback I think18:47
dhellmannah, ok18:47
dhellmannthis wasn't the only issue, we dealt with a keystone api change earlier in the day18:48
dhellmannso I'm thinking in the general, not specifically about this one thing, but about how we expose this class of problem before we're at a point where we're trying to branch18:48
dansmithmaster->master might've caught this specifically,18:48
dansmithbut we'd probably only want to run that periodically I'd think18:49
dhellmannI think it would have caught the keystone thing, too, because the issue there was the "new" side of grenade not getting users set up properly because of the keystone v3 difference18:49
dhellmannyeah, maybe weekly or something18:49
dansmithyeah18:49
dhellmannI'll put that on the retrospective notes for the release meetup at the ptg18:49
dhellmannhrm, we don't seem to have a shared release/qa session, that would be the ideal place for it18:50
*** mriedem1 has joined #openstack-release18:52
*** mriedem has quit IRC18:53
sdaguedhellmann: yeh, the master -> master job has been discussed in the past, there is some question about the amount of time it takes to handle that, vs. the fixup sets that we hit during cut over19:01
dhellmannsdague : it would be nice if this cut-over was a non-event19:02
sdaguedhellmann: sure, you just push the events other places19:02
sdagueand more likely paper over things19:03
dhellmannwell, sure. we either gate everything on not letting that job break or we have advance notice when there's less schedule pressure19:03
sdagueso, that keystone bug wouldn't exist if we hadn't tripped it here during cutover19:03
dhellmannwouldn't it have been caught by a master->master job?19:03
sdaguebecause on the devstack side these changes were made incrementally enough that I definitely didn't understand the implications of what stopped working19:03
sdaguedhellmann: nope, it would have been iterated through as "keystone v3 required support"19:04
*** mriedem1 is now known as mriedem19:04
dhellmannok, well, the responses on the bug so far seem to imply that's how the keystone team sees it19:04
sdagueyeh, it might be, but I think very few non keystone folks believe that :)19:05
dhellmannI don't really know.19:05
dhellmannat this rate, I don't expect to finish this work today, though, which makes it less likely we'll be ready to open pike this week as planned19:06
dhellmannso I'm looking for ways we could have avoided that -- I don't like these all-hands-on-deck situations19:06
dansmithsdague: mriedem: can I see the post-stack.sh that we generate anywhere?19:14
dansmithapparently I have a syntax error but I'm not seeing it19:14
mriedemdansmith: i dont think so19:16
dansmithalthough19:16
sdaguedansmith: it's going to be the escaping problem19:16
sdagueit's a heredoc in shell, with shell variables19:16
dansmithI shouldn't even be getting here19:16
sdaguemove the discover assignment outside of it19:16
dansmithokay19:17
sdaguedansmith: I can do it if you like, unless you've got the patch already19:18
dansmithit's already done19:19
dansmithany reason we don't capture post-stack.sh for logs?19:19
dansmithor just never needed it?19:19
sdaguehonestly, just never needed it19:20
sdagueit was one line of ansible previously19:20
dansmithyeah19:20
sdagueit's fine to do a thing to capture it, or even cat it out19:21
dansmithI didn't find where it got written in the logs either, but maybe it's somewhere19:21
sdaguewe might be under reduced logging at that point19:21
*** kgiusti has quit IRC19:31
*** kozhukalov has joined #openstack-release19:33
*** sridhar_ram has joined #openstack-release19:39
dansmithguh19:43
dhellmannwhat now?19:46
dhellmannoh, heh19:47
sdaguedansmith: heh, I was actually debugging at the same time, I just put it in an if clause instead19:48
sdagueI apparently pushed 3 minutes after you19:48
dansmithmkay19:49
dhellmanndansmith, sdague: should the whole script fail if the command line 649/650 fails?19:50
dansmithdhellmann: we're -e19:50
dhellmannah, ok, that wasn't clear19:50
dhellmannI see the -x, I guess the caller sets -e?19:50
sdagueyeh, the caller is running under -e19:51
*** ashtokolov has joined #openstack-release19:51
dhellmannk19:51
*** betherly has joined #openstack-release19:51
*** mwhahaha has joined #openstack-release19:53
*** sripriya has joined #openstack-release19:55
*** robcresswell has joined #openstack-release19:56
*** david-lyle has quit IRC19:57
*** igorbelikov_ has joined #openstack-release19:58
*** kong has joined #openstack-release19:59
*** spzala has joined #openstack-release19:59
*** spzala has quit IRC19:59
*** mriedem has quit IRC20:02
*** tommylikehu_ has joined #openstack-release20:02
*** nikhil has joined #openstack-release20:03
*** mriedem has joined #openstack-release20:04
*** electrofelix has quit IRC20:04
*** johnsom has joined #openstack-release20:05
*** tommylikehu_ has quit IRC20:06
*** dasanind has joined #openstack-release20:14
*** david-lyle has joined #openstack-release20:15
*** david-lyle has quit IRC20:20
*** spzala has joined #openstack-release20:21
*** spzala has quit IRC20:21
*** cody-somerville has joined #openstack-release20:25
*** cody-somerville has quit IRC20:25
*** cody-somerville has joined #openstack-release20:25
*** AJaeger has left #openstack-release20:26
*** kzaitsev_mb has joined #openstack-release20:29
sdaguedhellmann / dansmith looks like things are passing with the last set of patches20:32
dhellmann\o/20:32
*** amoralej is now known as amoralej|off20:39
*** pcaruana has quit IRC20:43
*** sdake has joined #openstack-release20:54
*** sdake_ has quit IRC20:58
dims_nice!20:58
*** wolverineav has quit IRC20:59
*** wolverineav has joined #openstack-release20:59
*** wolverineav has quit IRC20:59
*** wolverineav has joined #openstack-release21:01
openstackgerritDoug Hellmann proposed openstack/releases master: import propose-final-releases command from release_tools  https://review.openstack.org/43116421:04
openstackgerritDoug Hellmann proposed openstack/releases master: import a version of list_unreleased_changes.sh  https://review.openstack.org/43116521:04
openstackgerritDoug Hellmann proposed openstack/releases master: import list_library_unreleased_changes.sh  https://review.openstack.org/43116621:04
openstackgerritDoug Hellmann proposed openstack/releases master: teach list-deliverables to filter on governance tags  https://review.openstack.org/43116721:04
openstackgerritDoug Hellmann proposed openstack/releases master: import list_stable_unreleased_changes.sh from release-tools  https://review.openstack.org/43116821:04
openstackgerritDoug Hellmann proposed openstack-infra/release-tools master: remove list-repos command  https://review.openstack.org/43116921:04
openstackgerritDoug Hellmann proposed openstack-infra/release-tools master: move propose-final-releases command to releases repo  https://review.openstack.org/43117021:04
openstackgerritDoug Hellmann proposed openstack-infra/release-tools master: move list_unreleased_changes.sh to the releases repo  https://review.openstack.org/43117121:04
openstackgerritDoug Hellmann proposed openstack-infra/release-tools master: move list_library_unreleased_changes.sh to releases repo  https://review.openstack.org/43117221:04
openstackgerritDoug Hellmann proposed openstack-infra/release-tools master: remove list_oslo_unreleased_changes.sh  https://review.openstack.org/43117321:04
openstackgerritDoug Hellmann proposed openstack-infra/release-tools master: move list_stable_unreleased_changes.sh to releases repo  https://review.openstack.org/43117421:04
openstackgerritDoug Hellmann proposed openstack-infra/release-tools master: remove branching scripts  https://review.openstack.org/43117521:04
openstackgerritDoug Hellmann proposed openstack-infra/release-tools master: remove ptl command  https://review.openstack.org/43117621:04
openstackgerritDoug Hellmann proposed openstack-infra/release-tools master: remove highest_semver.py  https://review.openstack.org/43117721:04
openstackgerritDoug Hellmann proposed openstack-infra/release-tools master: remove latest-deliverable-versions command  https://review.openstack.org/43117821:04
*** cody-somerville has quit IRC21:09
*** askb has joined #openstack-release21:15
*** cody-somerville has joined #openstack-release21:16
*** cody-somerville has quit IRC21:16
*** cody-somerville has joined #openstack-release21:16
*** dave-mccowan has quit IRC21:26
openstackgerritAlex Schultz proposed openstack/releases master: Release secondary puppet modules  https://review.openstack.org/43093721:27
openstackgerritAlex Schultz proposed openstack/releases master: Release puppet-congress Ocata rc1  https://review.openstack.org/43118221:29
*** adrian_otto has quit IRC21:29
*** Jeffrey4l has quit IRC21:34
*** Jeffrey4l has joined #openstack-release21:34
openstackgerritAlex Schultz proposed openstack/releases master: Release secondary puppet modules  https://review.openstack.org/43093721:36
*** dave-mccowan has joined #openstack-release21:46
stevemarsdague: dhellmann dtroyer everything turn out okay? just logging back on now21:46
dhellmannstevemar : we're past that issue, still waiting for another fix to land21:46
*** adrian_otto has joined #openstack-release21:48
sdagueit looks like all the rest of it is in the gate now21:48
dhellmannyep21:48
dhellmannnext up is to get tonyb to unblock the xstatic version changes so we can merge that before branching the requirements repo21:49
openstackgerritFeilong Wang proposed openstack/releases master: Release ZaqarClient 1.4.0  https://review.openstack.org/43119421:53
*** sdake has quit IRC22:02
*** sdake has joined #openstack-release22:02
*** tommylikehu_ has joined #openstack-release22:02
*** tommylikehu_ has quit IRC22:07
stevemardhellmann: okay, thanks for the quick reply, sorry about being afk22:10
dhellmannnp, it sounded like you had something important to take care of.22:10
dims_dhellmann : you have been busy today :) https://review.openstack.org/#/q/status:open+project:openstack-infra/release-tools+branch:master+topic:move-tools-to-releases22:20
dims_are they just moving this around?22:20
dhellmanndims_ : yeah, that's based on the comments from ttx on that patch adding the new tools22:20
dhellmannalthough I haven't actually moved those new tools over, I should do that22:20
dhellmannthere's no rush to approve those; I don't want to break something we're going to use during the final release22:21
dhellmannbut it was a convenient thing to do while waiting for the d-g and grenade fixes to land22:21
dims_things around?22:21
dhellmannyeah, those are mostly scripts that use the data from the releases repo as input22:22
dhellmannor that call list-deliverables22:22
dims_ack dhellmann22:26
*** cody-somerville has quit IRC22:28
*** dave-mccowan has quit IRC22:31
*** sdake_ has joined #openstack-release22:32
*** sdake has quit IRC22:36
*** mriedem has quit IRC22:46
openstackgerritDoug Hellmann proposed openstack/releases master: script to report on unreleased changes of deliverables with no releases  https://review.openstack.org/43121222:48
openstackgerritDoug Hellmann proposed openstack/releases master: add a summary report for inactive projects  https://review.openstack.org/43121322:48
*** inc0 has quit IRC22:52
*** mburned is now known as mburned_out22:52
*** cody-somerville has joined #openstack-release22:54
*** wolverineav has quit IRC23:02
*** wolverineav has joined #openstack-release23:03
*** wolverineav has quit IRC23:09
*** wolverineav has joined #openstack-release23:11
*** cody-somerville has quit IRC23:22
*** cody-somerville has joined #openstack-release23:22
*** cody-somerville has quit IRC23:22
*** cody-somerville has joined #openstack-release23:22
*** wolverineav has quit IRC23:22
*** ajo has quit IRC23:26
*** ajo has joined #openstack-release23:27
*** zhurong has joined #openstack-release23:30
*** zhurong has quit IRC23:36
*** gordc has quit IRC23:46
« Tuesday, 2017-02-07 Index Thursday, 2017-02-09 »

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