| *** sdake has quit IRC | 00:41 | |
| *** sdake has joined #openstack-release | 01:57 | |
| *** mtreinish has quit IRC | 01:59 | |
| *** mtreinish has joined #openstack-release | 02:01 | |
| *** mriedem has quit IRC | 02:11 | |
| *** mriedem has joined #openstack-release | 02:11 | |
| *** mriedem has quit IRC | 03:17 | |
| *** sdake_ has joined #openstack-release | 03:21 | |
| *** sdake has quit IRC | 03:22 | |
| *** bswartz has quit IRC | 03:55 | |
| *** bswartz has joined #openstack-release | 03:56 | |
| prometheanfire | I've noticed that projects have stable/mitaka branches now, requirements only gets that when the release date is hit? | 04:26 |
|---|---|---|
| *** sdake_ is now known as sdake | 04:29 | |
| *** sdake has quit IRC | 05:16 | |
| *** sdake has joined #openstack-release | 05:18 | |
| *** sdake_ has joined #openstack-release | 05:21 | |
| *** sdake has quit IRC | 05:23 | |
| *** sdake_ is now known as sdake | 05:45 | |
| *** pcaruana has quit IRC | 05:48 | |
| *** bdemers has quit IRC | 06:03 | |
| *** bdemers has joined #openstack-release | 06:07 | |
| *** lifeless has quit IRC | 07:06 | |
| *** pcaruana has joined #openstack-release | 07:32 | |
| *** daemontool has joined #openstack-release | 07:33 | |
| ttx | prometheanfire: no, we should cut the requirements branch today | 07:44 |
| ttx | dhellmann: ^ I think today is better than tomorrow for that | 07:44 |
| *** sdake has quit IRC | 08:02 | |
| *** sdake has joined #openstack-release | 08:06 | |
| *** sdake has quit IRC | 08:18 | |
| *** kzaitsev_mb has joined #openstack-release | 09:02 | |
| *** vgridnev has joined #openstack-release | 09:05 | |
| openstackgerrit | Merged openstack/releases: Release ironic-ui 1.1.0 https://review.openstack.org/296609 | 09:27 |
| *** kzaitsev_mb has quit IRC | 09:27 | |
| *** sdague has joined #openstack-release | 10:24 | |
| *** sheeprine has quit IRC | 10:28 | |
| *** sheeprine has joined #openstack-release | 10:30 | |
| openstackgerrit | Sylvain Bauza proposed openstack/releases: Tag Nova Mitaka RC2 https://review.openstack.org/297006 | 10:40 |
| *** daemontool has quit IRC | 10:46 | |
| *** lifeless has joined #openstack-release | 11:07 | |
| odyssey4me | hmm, something odd has happened - OpenStack-Ansible has disappeared from the releases page | 11:24 |
| odyssey4me | ok, found where it is in the repo - it's been moved: https://github.com/openstack/releases/commit/5e15c8b6ab49486e242e23d3d028fbf7a85b9360 | 11:26 |
| odyssey4me | fair enough, but now where does it get published? | 11:26 |
| odyssey4me | ah, found it http://releases.openstack.org/independent.html#openstack-ansible | 11:27 |
| dims | odyssey4me : cool you found it | 11:28 |
| odyssey4me | :) yeah, thanks | 11:28 |
| *** _amrith_ is now known as amrith | 11:30 | |
| *** vgridnev has quit IRC | 11:36 | |
| *** vgridnev has joined #openstack-release | 11:37 | |
| *** daemontool has joined #openstack-release | 12:07 | |
| *** TravT has quit IRC | 12:09 | |
| *** daemontool has quit IRC | 12:15 | |
| *** mriedem has joined #openstack-release | 12:21 | |
| *** gordc has joined #openstack-release | 12:34 | |
| *** amrith is now known as _amrith_ | 12:52 | |
| dhellmann | ttx: yeah, we should go ahead and create the stable branch | 13:11 |
| dhellmann | ttx: should we tag too? I only see *-eol tags | 13:12 |
| ttx | no tag that I know of | 13:40 |
| ttx | dhellmann: ^ | 13:40 |
| ttx | please branch | 13:41 |
| * ttx has a look at the mitaka jobs | 13:41 | |
| dhellmann | ok, branching requirements now | 13:43 |
| dims | dhellmann : sdague and i have created stable/mitaka branches for devstack and grenade, added info to the etherpad | 13:44 |
| ttx | can't spot any utter failure on the stable-mitaka jobs that have supposedly run this morning | 13:44 |
| ttx | but then I have trouble finding them on the multiple-jenkinses | 13:44 |
| dhellmann | ok, stable/mitaka is created in the requirements repo and the gitreview update is in CI | 13:45 |
| dhellmann | https://review.openstack.org/297117 | 13:45 |
| dims | dhellmann : ooh and devstack-gate is branchless | 13:45 |
| dhellmann | ttx, dims: I'm going through the TODO list items in R-2 to mark anything we've done as completed | 13:47 |
| dims | dhellmann : ack, just took care of " QA Team to branch devstack / grenade for upgrade / gate testing" | 13:47 |
| dhellmann | are we ready to unfreeze requirements? | 13:48 |
| dims | dhellmann : i'd give it another day, need to check grenade mitaka jobs are good. | 13:49 |
| dims | sdague : ^^ | 13:50 |
| ttx | we usually do one day of observation | 13:50 |
| dhellmann | sounds good | 13:50 |
| ttx | usually after a few hours sdague comes and tells us the universe just broke | 13:50 |
| dhellmann | heh | 13:50 |
| dims | LOL | 13:50 |
| ttx | curious if we got it right this time | 13:50 |
| dims | dhellmann : whoa the bot already proposed patches https://review.openstack.org/#/q/is:open+branch:stable/mitaka | 13:50 |
| sdague | so, right now, things are looking pretty good | 13:51 |
| ttx | wat | 13:51 |
| dhellmann | a bunch of whitespace changes? | 13:51 |
| sdague | there was only 1 grenade bug that I needed to fix before the branch upgrades worked | 13:51 |
| dims | mriedem : how do you want to deal with bot proposed updates to stable/ branches? (note we have not unfrozen requirements yet) | 13:52 |
| ttx | dhellmann: more than whitespace | 13:52 |
| dhellmann | ah, the first one I clicked on was | 13:52 |
| dhellmann | grrr | 13:52 |
| *** daemontool has joined #openstack-release | 13:52 | |
| dhellmann | are these all for projects that never merged their updates in the first place? | 13:52 |
| ttx | why does the bot spot a difference ? Should be the same as master/master difference, no ? | 13:52 |
| ttx | swift does not merge updates, but that shall be the only "different" project | 13:53 |
| mriedem | dims: like requirements updates? | 13:53 |
| mriedem | for stable/mitaka? | 13:53 |
| dims | mriedem : https://review.openstack.org/#/q/owner:%22OpenStack+Proposal+Bot%22+status:open+branch:stable/mitaka | 13:53 |
| ttx | dhellmann: I thought we merged all requirements updates while in freeze | 13:53 |
| dims | mriedem : y | 13:53 |
| ttx | maybe we missed that step | 13:53 |
| dhellmann | ttx: perhaps not? | 13:53 |
| dhellmann | yeah | 13:54 |
| ttx | dhellmann: sad, since that is why we actually freeze | 13:54 |
| dhellmann | yeah | 13:54 |
| mriedem | dims: i'd say just do it the same as always | 13:54 |
| dhellmann | I wonder if there's some way to add that to the validation? if the list-changes job produced a list of unmerged requirements updates on the branch, for example | 13:54 |
| dims | mriedem : ack just making sure | 13:55 |
| ttx | dhellmann: all libs for some reason | 13:55 |
| dims | dhellmann : neutron too (not just libs) | 13:55 |
| dhellmann | right | 13:56 |
| ttx | + neutron yes | 13:56 |
| dhellmann | alright, well, now what? | 13:56 |
| dhellmann | should we merge these? | 13:56 |
| dims | mriedem : another question : do you want to switch off bot proposing new library versions to openstack/requirements in stable/mitaka | 13:56 |
| dhellmann | cut stable releases? | 13:56 |
| dhellmann | dims , mriedem : we want the automatic constraints updates turned off, but I think those only run in master already? | 13:57 |
| dims | dhellmann : i'll check | 13:57 |
| ttx | dhellmann: i'd like to understand why we are here first... Are those all (except neutron) libraries that did not have their requirements updates merges over the last 2 months ? | 13:58 |
| ttx | or at least not since lib freeze | 13:58 |
| dims | dhellmann : teams should evaluate these and propose releases if they really think they need it? | 13:58 |
| dims | ttx : oslo ones...since lib freeze | 13:58 |
| ttx | I'd say we need to cut them | 13:58 |
| dhellmann | dims : I would like to avoid the limbo situation we had with liberty requirements where no one felt comfortable approving | 13:58 |
| ttx | so that we are in a clean state | 13:59 |
| dims | dhellmann : right | 13:59 |
| dhellmann | let's get them all cleaned up -- right | 13:59 |
| dims | dhellmann : so we propose releases by ourselves? | 13:59 |
| dhellmann | dims : we'll need to coordinate with the lib owners, but yeah, something like that | 14:00 |
| dims | dhellmann : problem is some teams have already backported stuff (oslo.messaging) | 14:00 |
| dhellmann | that's ok, they did that intending to release it, right? | 14:00 |
| dims | so it's not just requirements changes | 14:00 |
| ttx | i think that's ok | 14:00 |
| dims | yep | 14:00 |
| dhellmann | it's still only going to get a patch version update | 14:00 |
| dims | right | 14:00 |
| ttx | think of it as last mitaka release for libs | 14:01 |
| dims | so we should tell people to review at least | 14:01 |
| ttx | (pre-release) | 14:01 |
| dims | +1 ttx | 14:01 |
| dhellmann | the oslotest change represents a change that was merged in master after stable/mitaka was created | 14:02 |
| dhellmann | the patch date for that was feb 26 | 14:02 |
| ttx | hm | 14:02 |
| dhellmann | so that at least, I think, is one where we cut a stable branch in the lib before we had merged the requirements updates? | 14:03 |
| dhellmann | I wonder if I shouldn't have used openstack/requirements head for the branch. | 14:03 |
| ttx | we should basically freeze requirements before we cut the branch for libs, and not cut the branch for libs until we have synced the reqs | 14:03 |
| dims | dhellmann : that's the only option | 14:03 |
| ttx | might be difficult with the earlier deadlines | 14:04 |
| dhellmann | we've had several updates to requirements in the last 2 weeks | 14:04 |
| dims | ttx : won't work because libs tend to break which we fix by changing requirements | 14:04 |
| dhellmann | http://git.openstack.org/cgit/openstack/requirements/log/ | 14:04 |
| dhellmann | all of those look legit, though | 14:05 |
| dims | right dhellmann | 14:05 |
| ttx | right, so those should be merged | 14:05 |
| ttx | None of those releases should trigger a requirement change, right | 14:06 |
| dims | right, unless something breaks the world | 14:06 |
| dhellmann | ttx: they need the constraints file updated | 14:06 |
| dhellmann | and right, what dims said | 14:06 |
| ttx | oh well, maybe that's an artifact of the process then | 14:07 |
| ttx | didn't expect so many | 14:07 |
| dhellmann | yeah | 14:07 |
| prometheanfire | ttx: thanks, I might cut the live ebuild for keystone today as well | 14:07 |
| ttx | let's try to communicate to PTLs and release as much today as we can, because Friday | 14:07 |
| stevemar | o/ | 14:07 |
| ttx | although I guess we have all week next week too | 14:07 |
| *** _amrith_ is now known as amrith | 14:07 | |
| ttx | dhellmann: maybe mark the ones that shall be updated on the spreadsheet | 14:08 |
| dhellmann | yeah, I'll start recording that | 14:08 |
| dhellmann | an easier list to review: https://review.openstack.org/#/q/is:open+branch:stable/mitaka+topic:openstack/requirements | 14:09 |
| *** sigmavirus24_awa is now known as sigmavirus24 | 14:09 | |
| dhellmann | ttx, dims, mestery, armax : there's already a neutron-lib 0.0.2 and we have requirements updates for stable/mitaka. I suppose we'll never release a stable update for that lib? | 14:15 |
| dims | dhellmann : armax : ttx : +1 to merge bot proposed update and re-propose the release with a new SHA | 14:16 |
| mestery | dhellmann: I think you're correct | 14:16 |
| dhellmann | dims : re-propose 0.0.2? we don't have a way to do that | 14:17 |
| dhellmann | dims , ttx: I've added "stable req update" as a special note in the dashboard for the projects that have them | 14:18 |
| ttx | dhellmann: we could also unbranch stable/mitaka there | 14:18 |
| ttx | dhellmann: I don't think anything uses neutron-lib yet ? | 14:18 |
| dims | dhellmann : ah i thought it was in review | 14:18 |
| ttx | i.e. consider 0.0.1 and 0.0.2 part on Newton | 14:18 |
| dhellmann | mestery , armax : ^^? | 14:18 |
| ttx | a 0.0.1 in stable with a 0.0.2 next is a dead tree anyway | 14:19 |
| ttx | so unless we REALLY need a 0.0.1 in stable/mitaka for some reason... | 14:19 |
| mestery | dhellmann: I'd like to make sure we pull dougwig into this discussion before we decide here. I think what you're saying makes sense, but I'd be more comfortable getting dougwig to +1 as well. | 14:19 |
| ttx | sure | 14:19 |
| mestery | cool | 14:19 |
| ttx | we already have a number of libs that won't have a mitaka release so there is precedent | 14:20 |
| ttx | (new libs that are not ready and not used) | 14:20 |
| dhellmann | ttx: do you have +2 on keystoneauth stable/mitaka? | 14:21 |
| ttx | I suspect not,; checking | 14:21 |
| stevemar | dhellmann: need something? | 14:21 |
| ttx | confirmed no | 14:21 |
| stevemar | dhellmann: this patch? https://review.openstack.org/#/c/297123/ | 14:22 |
| patchbot | stevemar: patch 297123 - keystoneauth (stable/mitaka) - Updated from global requirements | 14:22 |
| dhellmann | ttx: ok, I'll work on a project-config patch for that | 14:22 |
| dhellmann | stevemar : yeah, your acls on keystoneauth are wrong | 14:22 |
| stevemar | dhellmann: should i include a group? | 14:23 |
| dhellmann | stevemar : your stable branch should be locked down right now: https://review.openstack.org/297182 | 14:24 |
| dhellmann | ttx: ^^ | 14:24 |
| ttx | looking up acl | 14:24 |
| dhellmann | I copied those settings from keystonemiddleware | 14:25 |
| ttx | dhellmann: lgtm | 14:26 |
| ttx | wow that ironic-ui announce email looks completely wrong | 14:28 |
| ttx | dhellmann: seen that ? | 14:28 |
| ttx | http://lists.openstack.org/pipermail/openstack-announce/2016-March/001032.html | 14:28 |
| *** sdake has joined #openstack-release | 14:37 | |
| dhellmann | ugh, what is that? | 14:38 |
| bknudson | are we going to need releases for keystoneauth? | 14:38 |
| dhellmann | bknudson : yes | 14:38 |
| bknudson | dhellmann: minor or major version increase? | 14:38 |
| dhellmann | bknudson : tbd | 14:38 |
| bknudson | minor or patch I should have said. | 14:38 |
| dhellmann | ttx: that come out of the job? it looks like maybe? | 14:39 |
| dims | ouch on the email! | 14:39 |
| dhellmann | ttx: or was that the one we did by hand yesterday because the job wasn't there? | 14:40 |
| dhellmann | I don't see an announce job in http://logs.openstack.org/9a/9ae8de43c21ddc1fc1c76e68b9b2daabb8e4b60d/release/ | 14:40 |
| ttx | dhellmann: no, job output | 14:40 |
| ttx | sent by no-reply and all | 14:40 |
| ttx | would be good to reply to it with the correct version | 14:41 |
| dims | dhellmann "{{ description }}" went bad http://git.openstack.org/cgit/openstack-infra/release-tools/tree/releasetools/release_notes.py#n74 | 14:41 |
| dhellmann | oh, I'm looking at the wrong repo | 14:41 |
| ttx | Today is a bit crazy for me + my wife is sick so I have to use every extra second to care for the kids | 14:41 |
| dhellmann | ok, I'll look into this | 14:42 |
| ttx | thanks! | 14:42 |
| dhellmann | ttx: before you step away, I was going to compose email about the stable requirements updates, I'd like you to review it before I send | 14:42 |
| ttx | sure | 14:42 |
| ttx | just ping me when ready | 14:42 |
| dhellmann | ok | 14:43 |
| ttx | I'm not really stepping away, just avoiding everything I can | 14:43 |
| dims | ttx : ouch, yes, i am around to help | 14:43 |
| ttx | evasion maneuvers | 14:43 |
| dhellmann | ttx, dims: https://etherpad.openstack.org/p/ELLLuErGuN | 14:47 |
| dims | dhellmann : do we tell them to file reviews with version # and SHA? | 14:52 |
| dims | or do we do it for them? | 14:52 |
| dims | that bit was not clear | 14:52 |
| dhellmann | dims : yeah, I should be more explicit about that | 14:52 |
| dims | yes please | 14:52 |
| dhellmann | mriedem, dims, ttx: last time we did requirements updates in a stable branch we used patch versions. Should we do that here, or go ahead with minimum version updates since we haven't started releasing from master? | 14:54 |
| dhellmann | bknudson : ^^ | 14:54 |
| bknudson | y, since we haven't released from master yet we can do whatever we want | 14:54 |
| bknudson | whatever we do here just affects what we do when we release master. | 14:55 |
| ttx | dhellmann: reviewing | 14:55 |
| bknudson | I would prefer minor version #s here since it's reqs update. | 14:55 |
| dhellmann | ttx, jroll : ironic-ui has python-ironicclient>=1.1 in setup.py, which means it installs that when we run "python setup.py --description" so that's why the announce email contains all of that garbage | 14:56 |
| ttx | hmm | 14:56 |
| dhellmann | jroll : do you have any idea why that dependency is in setup.py instead of requirements.txt? | 14:56 |
| jroll | dhellmann: ouch :( | 14:56 |
| jroll | I don't - betherly ? | 14:56 |
| dhellmann | it's in both, actually | 14:57 |
| ttx | I guess min update is an option there | 14:57 |
| ttx | you should spell it out as X.Y+1.0 since so many people have issues with semver | 14:58 |
| dhellmann | ok, let's go ahead with a minimum version | 14:58 |
| dhellmann | yeah | 14:58 |
| dims | +1 for X.Y+1 | 14:59 |
| dhellmann | how about: X.Y.0 + 0.1.0 | 14:59 |
| ttx | not sure that's clearer tbh | 14:59 |
| dhellmann | "increment the Y from X.Y.Z and reset Z to 0" | 15:00 |
| dims | works dhellmann | 15:00 |
| ttx | works | 15:00 |
| dhellmann | ok, sending now | 15:00 |
| ttx | maybe wait until lifeless is up before approving them so that he has a chance to chime in | 15:01 |
| dhellmann | approving the releases? | 15:02 |
| dhellmann | or the requirements updates? | 15:02 |
| mriedem | dhellmann: i think we do minor version updates on stable branches if the requirements changes are higher minimum required versions of a dependency | 15:03 |
| dhellmann | mriedem : some of these are | 15:03 |
| ttx | I'll try to find some time to process nova rc2 in a few | 15:04 |
| mriedem | dhellmann: dims: re: " we want the automatic constraints updates turned off, but I think those only run in master already?" i think we want the u-c bot proposed updates disabled for stable/mitaka if they aren't already | 15:04 |
| dhellmann | mriedem : right, dims looked at that but if he reported back I missed what he said | 15:04 |
| dhellmann | ttx: thanks | 15:04 |
| mriedem | dhellmann: dims: yeah, sorry, was in the nova meeting last hour | 15:04 |
| dhellmann | mriedem : np, that was a passive request for dims to repeat what he said earlier :-) | 15:05 |
| *** lifeless has quit IRC | 15:13 | |
| *** david-lyle has quit IRC | 15:13 | |
| *** cody-somerville has quit IRC | 15:13 | |
| *** skraynev has quit IRC | 15:13 | |
| dims | dhellmann : mriedem : got side-tracked | 15:14 |
| dims | dhellmann : mriedem : looking now | 15:14 |
| gordc | dhellmann: just a headsup, we'll need a ceilometer rc2 (i forgot to remove some entry_points) | 15:15 |
| gordc | is there a date i should request rc2 by? | 15:15 |
| *** kzaitsev_mb has joined #openstack-release | 15:17 | |
| *** david-lyle has joined #openstack-release | 15:17 | |
| dhellmann | jroll, betherly : https://bugs.launchpad.net/ironic-ui/+bug/1561581 | 15:18 |
| openstack | Launchpad bug 1561581 in Ironic UI "packaging instructions include extra dependencies" [Undecided,New] | 15:18 |
| dhellmann | gordc : next thursday: http://releases.openstack.org/mitaka/schedule.html | 15:18 |
| *** armax has quit IRC | 15:19 | |
| gordc | dhellmann: ack. i'll probably request it monday then to minimise rc requests | 15:20 |
| *** armax has joined #openstack-release | 15:20 | |
| dhellmann | gordc : ok | 15:21 |
| ttx | processing nova rc2 | 15:23 |
| *** sdake_ has joined #openstack-release | 15:25 | |
| *** sdake has quit IRC | 15:25 | |
| betherly | dhellmann: jroll I will definitely take a look at it and check it runs without being in setup.py. Block booked on stuff today but will look Tuesday and ping you all re what I find | 15:26 |
| dhellmann | betherly : I'm testing a patch locally now, stand by | 15:26 |
| sdake_ | dhellmann there is a python dependnecy problem from python-extras and soemthing related to pbr on a fresh system | 15:26 |
| sdake_ | http://paste.openstack.org/raw/491729/ | 15:26 |
| betherly | dhellmann: oh ok thanks | 15:26 |
| sdake_ | dhellmann because of cisco policy, I am not permitted to submit changes outside of the openstack ecosystem without alot of hassle | 15:27 |
| dhellmann | sdake_ : "python-extras"? | 15:27 |
| sdake_ | dhellmann ya the obvious solution is to install python-extras | 15:27 |
| sdake_ | but i think the root of the problem is a dep is unaccounted for in a requirements.txt file | 15:27 |
| dhellmann | sdake_ : I don't know what that is | 15:27 |
| sdake_ | pbr imports the chain of dependencies that end up using python extras | 15:28 |
| clarkb | sdake_: dhellmann looks like missing python dev headers? | 15:28 |
| sdake_ | it can be obtained by yum intall python-extras | 15:28 |
| sdake_ | i am nto sure what the pip equpvalent is | 15:28 |
| sdake_ | clarkb yes he is definately missing xcode | 15:28 |
| sdake_ | i just think its appropriate to get the requirements.txt files for whatever project is broken fixed | 15:29 |
| sdake_ | but as I mentioned - cisco policy prevents that - or makes my life reallly difficult to do so | 15:29 |
| clarkb | the fix is add dev headers I think | 15:29 |
| *** lifeless has joined #openstack-release | 15:29 | |
| *** cody-somerville has joined #openstack-release | 15:29 | |
| *** skraynev has joined #openstack-release | 15:29 | |
| sdake_ | clarkb adding xcode may just install extras | 15:29 |
| *** lifeless has quit IRC | 15:30 | |
| *** lifeless has joined #openstack-release | 15:30 | |
| jroll | dhellmann: betherly: thanks | 15:30 |
| jroll | sdake_: clarkb: looks like just a missing python module, no? https://github.com/testing-cabal/subunit/blob/master/setup.py#L13 | 15:30 |
| jroll | https://pypi.python.org/pypi/extras | 15:31 |
| clarkb | yes because installcompilation failed? | 15:31 |
| sdake_ | jroll yes i agree, suvvmit is the problem | 15:32 |
| sdake_ | but it looks like it doesn't hae requirements.txt files | 15:32 |
| jroll | clarkb: looks like a warning to me | 15:32 |
| sdake_ | clarkb that is a warning - it says it uses python code but its slower instead | 15:32 |
| jroll | simplejson optimized if headers are available | 15:32 |
| dhellmann | jroll , betherly : https://review.openstack.org/297219 | 15:32 |
| clarkb | oh its warlock failing a few lines below simplejson | 15:32 |
| jroll | dhellmann: lgtm | 15:33 |
| sdake_ | the poitn is there is an unaccounted for dependency in subuit it appears, or its not getting installed via tox | 15:33 |
| sdake_ | should i add it to kolla's requirements instead, or what is the correct fix exactly? | 15:33 |
| dhellmann | jroll , betherly : stable/mitaka backport: https://review.openstack.org/#/c/297220/ | 15:34 |
| patchbot | dhellmann: patch 297220 - ironic-ui (stable/mitaka) - Remove python-ironicclient from setup_requires | 15:34 |
| openstackgerrit | Merged openstack/releases: Tag Nova Mitaka RC2 https://review.openstack.org/297006 | 15:34 |
| sdake_ | jroll any thoughts? | 15:35 |
| dhellmann | sdake_ : it sounds like what clarkb is saying is warlock has the bad dependency, so I would think we should fix it there | 15:35 |
| jroll | sdake_: it's in subunit's install_requires, which I think should be sufficient but maybe not? https://github.com/testing-cabal/subunit/blob/master/setup.py#L13 | 15:35 |
| sdake_ | dhellmann warlock just runs unopitimized without python development headers | 15:35 |
| dhellmann | hmm, no, it's not warlock. I see "positional" | 15:36 |
| clarkb | extras shows up in the list | 15:36 |
| dhellmann | line 316 in http://paste.openstack.org/show/491729/ | 15:37 |
| clarkb | the log isnt exactly the clearest of things. It is running setup.py in some project and that fails to import extras | 15:37 |
| sdake_ | more data, this isa fresh centos install with pip install kolla/ | 15:37 |
| sdake_ | followed by tox -e releasenotes | 15:37 |
| sdake_ | a third party is running the code to generate release notes in his env | 15:37 |
| sdake_ | in my env, it works fine | 15:37 |
| sdake_ | in his - its broken | 15:37 |
| sdake_ | dhellmann your right - its positional | 15:38 |
| sdake_ | 316 ftw | 15:38 |
| dhellmann | sdake_ : looks like that's owned by morgan | 15:38 |
| betherly | Thanks for sorting them dhellmann | 15:38 |
| betherly | workflowed | 15:38 |
| dhellmann | betherly : thanks, we should plan on another mitaka release for you when that's done | 15:39 |
| *** morgan has joined #openstack-release | 15:39 | |
| morgan | o/ | 15:39 |
| sdake_ | dhellmann but line 368 iports subunit which is the project which actually uses the dependency | 15:39 |
| sdake_ | hey morgan | 15:39 |
| sdake_ | just trying to sort out a missing dependency | 15:40 |
| dhellmann | morgan : line 316 of http://paste.openstack.org/show/491729/ starts an error reported trying to pip install kolla on a clean system | 15:40 |
| sdake_ | in one of the dependencies we use | 15:40 |
| dims | dhellmann : mriedem : https://review.openstack.org/297225/ (proposal bot) | 15:40 |
| dhellmann | sdake_ : you're right, maybe it really is subunit | 15:40 |
| betherly | dhellmann: when they've merged do I need to submit a release patch for 1.2? | 15:40 |
| sdake_ | not pip install kolla | 15:40 |
| sdake_ | tox -e releasenotes | 15:40 |
| dhellmann | betherly : yes, and see also my email to -dev about the requirements updates in case you have one of those patches (we should release both together) | 15:40 |
| dhellmann | sdake_ : oops, sorry | 15:40 |
| morgan | dhellmann: weird | 15:41 |
| clarkb | I think it is executing code at install time that shouldnt run until post install? | 15:41 |
| clarkb | so the extras module cant be found yet | 15:41 |
| dhellmann | clarkb : yeah, pip install testrepository gives me extras | 15:42 |
| sdake_ | hmm maybe testrepository is a missing dependency from kolla | 15:42 |
| sdake_ | should that be in kolla? :) | 15:42 |
| morgan | i mean, positional does pbr things... | 15:43 |
| dhellmann | morgan : oh, does it? | 15:43 |
| morgan | dhellmann: yes. it uses pbr | 15:43 |
| morgan | or it should be using it | 15:43 |
| morgan | the code is extremely simple | 15:43 |
| dhellmann | I see it using pbr in setup.py, is that what you mean? | 15:43 |
| dhellmann | that shouldn't cause this | 15:44 |
| morgan | yeah. | 15:44 |
| morgan | positional is insanely simple | 15:44 |
| sdake_ | kolla has testrepository | 15:44 |
| clarkb | ok how old is the tox/virtualenv in use? | 15:44 |
| morgan | dhellmann: i don't know how it could be causing an issue unless there is something wonky with pbr | 15:44 |
| clarkb | could be a bad old setupyools.behavior | 15:44 |
| dhellmann | clarkb : good call | 15:45 |
| sdake_ | clarkb i am not sure | 15:45 |
| dhellmann | sdake_ : find out if they installed setuptools from a system package, or from pip | 15:45 |
| *** sbezverk has joined #openstack-release | 15:45 | |
| sdake_ | clarkb i am haing sbezverk join the channel now | 15:45 |
| sbezverk | morning folks | 15:45 |
| dhellmann | pip, pbr, tox, virtualenv all need to be current and not system packages (mostly because those are rarely current) | 15:45 |
| clarkb | dhellmann: sdake_ tox depends on virtualenv which bundles setuptools | 15:45 |
| sdake_ | sbezverk run rpm -qi setuptools \ fpaste | 15:45 |
| sdake_ | or python-setuptoools | 15:45 |
| dhellmann | well, if you're installing setuptools from rpm that's probably your issue right there | 15:46 |
| sbezverk | FYI I am on ubuntu | 15:46 |
| dhellmann | the debs there are also usually out of date -- we recommend developers install python dev tools using pip | 15:46 |
| sdake_ | sbezverk ok well tell us if you using packaged version of setup tools or pip installed version | 15:46 |
| *** doug-fish has joined #openstack-release | 15:47 | |
| sdake_ | folks thanks for all the debug - my main concern is I wan tto make sure there isn't a missing deependency somewhere in our dep chain | 15:47 |
| ttx | dhellmann: I propose we just reply to the RC1 email to announce RC2s, rather than start a new thread | 15:47 |
| dhellmann | ttx: makes sense | 15:47 |
| dhellmann | sdake_ : sure. We have pretty minimal images in our gate, specifically for reasons like this. | 15:48 |
| sbezverk | http://paste.openstack.org/show/491745/ | 15:48 |
| sdake_ | dhellmann fwiw the gate works | 15:48 |
| dhellmann | right, that's what I mean. It's working in the gate, and on your system and mine, so I think this is an issue with sbezverk's system and not the packaging instructions | 15:48 |
| sdake_ | dhellmann got it makes sense | 15:49 |
| sbezverk | dhellmann very possible, I pasted the list of installed modules | 15:49 |
| dhellmann | sbezverk : latest tox is 2.3.1, latest setuptools is 20.3.1, so I'd try upgrading those | 15:49 |
| sdake_ | sbezverk aey ou using apt installed setuptoools | 15:49 |
| sbezverk | could you identify which one is not supported | 15:49 |
| clarkb | sbezverk: dhellmann specifically oyu have to upgrade virtualenv for tox to use newer setuptools since virtualenv bundles setuptools | 15:50 |
| dhellmann | sbezverk : the general rule is, when you are doing python development for openstack projects do not install the tools for that from system packages. Use get-pip.py to install pip, then use pip to install the other tools you need | 15:50 |
| dhellmann | clarkb : yeah, thanks, I missed that one | 15:50 |
| dhellmann | sbezverk : the latest virtualenv is 15.0.1 | 15:51 |
| dhellmann | so you can see that lots of those packages are *far* out of date | 15:51 |
| sdake_ | dhellmann ++ thanks :) | 15:52 |
| dhellmann | sdake_ : clarkb figured it out :-) | 15:52 |
| sdake_ | clarkb ++ :) | 15:52 |
| sdake_ | sbezverk do you need further assistance in getting your system updated? | 15:54 |
| sdake_ | if so, lets go over to #kolla and go through it there | 15:54 |
| sdake_ | apologies all for the idea that this was an unaccounted for dependency - i thought sbezverk was on centos where all this stuf fworks with system packaging :) | 15:55 |
| dhellmann | sdake_ : np, that's certainly what it looked like | 15:56 |
| sbezverk | updating all these modules broke pip somehow.. Is there a list with precise required version of required components | 15:56 |
| sbezverk | I will try to recover it offline | 15:56 |
| sdake_ | use the get-pip shell tool | 15:56 |
| *** palendae has joined #openstack-release | 16:12 | |
| *** mriedem is now known as mriedem_snow | 16:22 | |
| palendae | I have a question regarding reno and projects like openstack-ansible, which have their own subprojects - when submitting patches to the individual role repositories, release notes currently have to be duplicated in order to show up in both OSA and the role repository. Adding roll up functionality to reno would help with this, but I'm wondering if that's generally desirable for others | 16:23 |
| *** sheeprine has quit IRC | 16:25 | |
| *** vgridnev has quit IRC | 16:25 | |
| sbezverk | dhellmann I am still having an issue generating release notes, please see this paste: http://paste.openstack.org/show/491756/ | 16:26 |
| sdake_ | hey folks sbezverk still having same problem after upgrading all his deps | 16:26 |
| clarkb | sdake_: sbezverk make sure you run tox -r -e whatever | 16:26 |
| sbezverk | I have updated every component in pip | 16:26 |
| clarkb | to rebuild the virtualenv | 16:26 |
| sdake_ | clarkb thanks i dind't think of that | 16:26 |
| sbezverk | clarkb thank you, will try that now | 16:27 |
| sdake_ | sbezverk two options rm -rf .tox or tox -r -e releasenotes | 16:27 |
| dims | dhellmann : am done with my meetings for the day, reading scroll back | 16:28 |
| *** kzaitsev_mb has quit IRC | 16:28 | |
| *** vgridnev has joined #openstack-release | 16:28 | |
| sbezverk | clakrb same issue, see this paste: http://paste.openstack.org/show/491757/ | 16:29 |
| sdake_ | sbezverk can we get a pip show to validate your deps are all up to date | 16:30 |
| *** sheeprine has joined #openstack-release | 16:30 | |
| sdake_ | sbezverk i am not python wizard enough to knw, but folks in this channel can give a qiuck eval | 16:30 |
| sbezverk | sdake_ do you mean pip list? | 16:30 |
| sdake_ | pip show should show everything iI think | 16:30 |
| sdake_ | but maybe its pip list | 16:30 |
| clarkb | yes pip list | 16:30 |
| sdake_ | i'm ony at 8000 hours of experience with python | 16:30 |
| clarkb | that will include setuptools which freeze does not | 16:31 |
| sdake_ | so not quite expert level yet | 16:31 |
| sbezverk | sdake_ here you go http://paste.openstack.org/show/491758/ | 16:31 |
| sdake_ | maybel ess, like 6k hours | 16:31 |
| sdake_ | clarkb any ideas on the above paste? | 16:32 |
| sdake_ | six is definately old | 16:32 |
| sdake_ | pbr may be old | 16:32 |
| sdake_ | requests is s uper old | 16:32 |
| clarkb | pbr looks up to date | 16:32 |
| sdake_ | oh- testtools | 16:32 |
| clarkb | as do setuptools virtualenv and pip | 16:32 |
| sdake_ | testtools could be old no? | 16:33 |
| sdake_ | is extras only at 0.0.3? | 16:34 |
| *** sdake_ is now known as sdake | 16:34 | |
| clarkb | ya but thats not in the virtualenv so we only really care about the bootstrap a virtualenv code | 16:34 |
| sdake | sbezverk do this - rm -rf .tox in the kolla directory | 16:35 |
| clarkb | the other possibility is stale pbr in the repo itself, you might try with a clean checkout if not already | 16:35 |
| sbezverk | sdake I think you are looking at wrong place | 16:35 |
| sbezverk | log said that some .h files are missing | 16:35 |
| sdake | ok well that is an otpimization, but lets fix that | 16:36 |
| sdake | insall the python-dev package via apt | 16:36 |
| sbezverk | sdake done it now rerunning tox | 16:36 |
| *** vgridnev has quit IRC | 16:36 | |
| sbezverk | sdake it is better but still errors, now it did not find yaml.h | 16:37 |
| clarkb | fwiw I just ran that tox command on my local ubuntu machine and had no issues | 16:37 |
| sbezverk | do you know from the top of your head which dev tools provide this header? | 16:37 |
| clarkb | sbezverk: python-dev and libyaml-dev | 16:37 |
| sdake | yaml-dev possly? | 16:37 |
| clarkb | but those were listed as warnings so we ignored them initially, though its possible that they claim to be warnings but result in fails | 16:38 |
| sdake | clarkb yup might as well fix em ;) | 16:38 |
| sbezverk | clarkb sdake still fails different error though http://paste.openstack.org/show/491759/ | 16:41 |
| clarkb | that doesn't look like complete output? | 16:42 |
| clarkb | some of the compiles can be very noisy :/ | 16:42 |
| sdake | sbezverk where is th full log please | 16:42 |
| sdake | sbezverk don't forget to remove .tox | 16:43 |
| sbezverk | sdake clarkb the log gets cut by paste site. do you want me to email it to you? | 16:44 |
| sdake | pate on fedorapeopel.org | 16:45 |
| sdake | or pastebin | 16:45 |
| sdake | parather paste.fedoraproject.org | 16:46 |
| sbezverk | sdake http://paste.fedoraproject.org/344714/38018145/ | 16:47 |
| clarkb | its the same error at the end, no module named extras | 16:48 |
| clarkb | I have no idea honestly | 16:48 |
| sdake | sbezverk cd /usr/lib/pythoon*/site-packages | 16:49 |
| sdake | ls -ld extra* | paste | 16:50 |
| sbezverk | sdake I do not have site-packages folder | 16:52 |
| sdake | clarkb what i'm thinking is he is two extrs packages installed one distro and one pip | 16:52 |
| sdake | i dont know layout of ubuntu becausei use centos for dev | 16:52 |
| sdake | any tips on where to look? | 16:52 |
| clarkb | that shouldn't matter due to the virtualenv | 16:53 |
| clarkb | unless kolla allows site packages, let me look | 16:53 |
| clarkb | doesn't appear to | 16:53 |
| sbezverk | sdake I found one here: /nobackup/sbezverk/iscsi/kolla/.tox/releasenotes/lib/python2.7/site-packages | 16:53 |
| sdake | is there an extrs package in that directory? | 16:54 |
| *** odyssey4me has quit IRC | 16:55 | |
| sbezverk | sdake nope | 16:55 |
| *** odyssey4me has joined #openstack-release | 16:55 | |
| sdake | sbezverk add extrs to test-requriemetns.txt in kola | 16:56 |
| sdake | extras | 16:56 |
| sdake | this isn't a proper solution but i wan tto see if it makes it work | 16:56 |
| sbezverk | sdake found extras here /usr/lib/python2.7/dist-packages/extras | 16:56 |
| sdake | rm .tox as well | 16:56 |
| dhellmann | hey, guys, could you debug this in another channel, please? | 16:57 |
| sdake | dhellmann will do | 16:57 |
| sdake | sbezverk back to kolla | 16:57 |
| dhellmann | thanks, sorry | 16:57 |
| sdake | dhsorry for noise | 16:57 |
| sdake | dhellmann no reason to apoologize | 16:57 |
| sdake | i originally thought it was a dep problem in our dep chain but i think its just sbezverk environmental | 16:57 |
| sdake | dhellmann hence the escalation to this channel | 16:57 |
| dhellmann | yep, of course, np there | 16:58 |
| *** sridhar_ram has joined #openstack-release | 17:00 | |
| *** openstackgerrit has quit IRC | 17:01 | |
| *** openstackgerrit has joined #openstack-release | 17:01 | |
| odyssey4me | dhellmann back to palendae's question | 17:02 |
| *** kzaitsev_mb has joined #openstack-release | 17:02 | |
| odyssey4me | is there a way, right now, that we'd be able to do some sort of release note roll-up between repositories? | 17:02 |
| odyssey4me | repo A contains docs, release notes and code. repo A publishes those. repo A consumes repo B. repo B has release notes (kinda like a library). Could we work out a way for repo A to consolidate release notes from repo B and present them as a uniform set? | 17:04 |
| ttx | dhellmann: ok, Nova RC2 posted, see template for RC2 announces there | 17:19 |
| ttx | Hopefully kept the msgid and threads properly | 17:19 |
| *** gordc has quit IRC | 17:29 | |
| ttx | quick review of stable/mitaka changes before closing crtazyday | 17:30 |
| *** vgridnev has joined #openstack-release | 17:31 | |
| ttx | so far no violation of the "merge in master first" rule | 17:33 |
| ttx | alright, calling it a day | 17:34 |
| openstackgerrit | Armando Migliaccio proposed openstack/releases: Neutron Mitaka RC2 deliverables https://review.openstack.org/295398 | 17:34 |
| clarkb | grenade is on its way to understanding mitaka as well | 17:35 |
| *** sridhar_ram has quit IRC | 17:35 | |
| *** sridhar_ram has joined #openstack-release | 17:36 | |
| dhellmann | ttx: thanks | 17:38 |
| *** pcaruana has quit IRC | 17:40 | |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/releases: Release OpenStack-Ansible Kilo 11.2.12 / Liberty 12.0.9 https://review.openstack.org/297274 | 17:49 |
| *** kzaitsev_mb has quit IRC | 17:50 | |
| odyssey4me | dhellmann heh, that one's fun - it would appear that two releases at once doesn't work so well :) let me split them up | 18:04 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/releases: Release OpenStack-Ansible Kilo 11.2.12 https://review.openstack.org/297274 | 18:06 |
| *** sdake_ has joined #openstack-release | 18:07 | |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/releases: Release OpenStack-Ansible Liberty 12.0.9 https://review.openstack.org/297285 | 18:07 |
| dhellmann | odyssey4me : did it fail? as long as there aren't multiple releases in the same file it should work | 18:08 |
| dhellmann | oh, independent, yeah, that won't work, we'll have to merge one and then recheck the other | 18:09 |
| odyssey4me | dhellmann yeah, it was two in the same file - so the analysis output was only for one of them | 18:09 |
| dhellmann | yeah, it skips to the last one | 18:09 |
| odyssey4me | I've split them up - and they won't conflict with each other (different place in the file) | 18:09 |
| *** sdake has quit IRC | 18:10 | |
| dhellmann | did you submit them as a series? | 18:13 |
| odyssey4me | dhellmann nope - independent patches | 18:13 |
| dhellmann | ok. I'm not sure inserting something into the middle of the file actually works | 18:13 |
| odyssey4me | oh dear | 18:14 |
| dhellmann | we should make it, but I don't think that's a case we've dealt with yet | 18:14 |
| odyssey4me | originally we were in the series thing, but got moved to independent | 18:14 |
| dhellmann | yeah, ttx and I discussed that because of your release model | 18:14 |
| dhellmann | we might need to rethink that | 18:14 |
| odyssey4me | we really should be cycle-with-intermediary, but didn't make the cut-off date for requesting the change of release model | 18:14 |
| dhellmann | yeah | 18:15 |
| odyssey4me | I'm happy to submit a request to governance to change it if that would help? | 18:15 |
| odyssey4me | Although as I recall there are specific windows to do that. | 18:15 |
| dhellmann | yeah, go ahead and do that, and I'll revert the change that merged the files and then we can apply your releases on top of that change | 18:15 |
| odyssey4me | ok cool | 18:16 |
| dhellmann | odyssey4me : give me the link to the governance patch when you have it, so I can put it in the commit message for the revert | 18:16 |
| odyssey4me | will do | 18:16 |
| *** mriedem_snow is now known as mriedem | 18:20 | |
| odyssey4me | dhellmann does that make sense? https://review.openstack.org/297297 | 18:32 |
| *** daemontool has quit IRC | 18:32 | |
| openstackgerrit | Doug Hellmann proposed openstack/releases: Revert "move openstack-ansible under the _independent directory" https://review.openstack.org/297298 | 18:33 |
| dhellmann | odyssey4me : yes | 18:33 |
| odyssey4me | awesome, thanks | 18:34 |
| dims | dhellmann : we don't need to wait for governance model changes? | 18:35 |
| dhellmann | odyssey4me : I'll wait and let dims and ttx have a chance to look at both of our patches before continuing, but if you want to submit your new tags as follow-ups to the revert that would speed things along | 18:35 |
| dims | y was just reading both dhellmann | 18:35 |
| dhellmann | dims : no, the review timeline for those is pretty long (a week for lazy consensus) | 18:35 |
| dims | right. | 18:35 |
| odyssey4me | ok, let me get on that so long | 18:35 |
| dhellmann | the intent is there, as long as we don't downvote the governance change it's fine | 18:36 |
| *** gordc has joined #openstack-release | 18:36 | |
| odyssey4me | I expect that if the governance change is downvoted, then the release changes can always be resolved. | 18:36 |
| dims | odyssey4me : +2 | 18:36 |
| dhellmann | odyssey4me : yeah, we'll just have to make the tools smarter :-) | 18:38 |
| palendae | Speaking of smarter tools :) | 18:38 |
| palendae | Unless I'm misreading the backlog, have you considered the question about release notes from subprojects being rolled/bubbled up into a parent project, dhellmann? | 18:39 |
| dhellmann | palendae : that's not something we've considered. what's the use case? | 18:40 |
| *** vgridnev has quit IRC | 18:40 | |
| *** sridhar_ram1 has joined #openstack-release | 18:40 | |
| palendae | I'll steal odyssey4me's wording from above: repo A contains docs, release notes and code. repo A publishes those. repo A consumes repo B. repo B has release notes (kinda like a library). Could we work out a way for repo A to consolidate release notes from repo B and present them as a uniform set? | 18:40 |
| palendae | In this case, Repo A is openstack-ansible, repo B is a role that it pulls in (http://git.openstack.org/cgit/?q=openstack-ansible) | 18:41 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/releases: Release OpenStack-Ansible Kilo 11.2.12 https://review.openstack.org/297274 | 18:41 |
| *** sridhar_ram has quit IRC | 18:41 | |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/releases: Release OpenStack-Ansible Liberty 12.0.9 https://review.openstack.org/297285 | 18:42 |
| dhellmann | palendae : one challenge, maybe not a blocker, is that reno uses the source repo to determine what version # to associate with a note | 18:44 |
| palendae | So it would be the version number/tag of the repository the note is in? | 18:46 |
| dhellmann | that's what you would want, right? | 18:46 |
| dhellmann | you would want the collected notes to say things like "role X version A.B.C has note foo" | 18:46 |
| palendae | I do think that makes sense, I don't know if it's what others would want :) But it makes sense for me to not only denote the release notes at the OSA level, but the versions included | 18:47 |
| palendae | So basically OSA becomes an aggregate | 18:47 |
| dhellmann | how would you different role versions be organized in the OSA release notes? by OSA version of some sort? | 18:48 |
| palendae | I think a release of OSA would include 1-and-only-1 version of the role | 18:49 |
| dhellmann | how does reno know what that version is? | 18:49 |
| dhellmann | for that matter, how does anything else know? | 18:49 |
| palendae | As I understand odyssey4me's plan, that'll be pinned in https://github.com/openstack/openstack-ansible/blob/master/ansible-role-requirements.yml | 18:50 |
| palendae | Which I realize reno doesn't consume now | 18:50 |
| dhellmann | yeah, something could use that to drive reno though | 18:50 |
| odyssey4me | yeah, so the integrated release will have a SHA pinned relationship with the role. | 18:50 |
| odyssey4me | each role is consumable on its own - so it makes sense for it to have its own release notes | 18:51 |
| palendae | I think there was a larger discussion regarding the role versioning, which unfortunately I haven't been too involved with. That output's probably related to this | 18:51 |
| odyssey4me | OSA ideally would aggregate the release notes, but the release notes would be organised by the integrated release tags for that aggregated set | 18:51 |
| odyssey4me | if that seems plausible enough to do, I'd appreciate it if you could coach palendae enough to work out a PoC | 18:52 |
| palendae | I'm guessing reno itself wouldn't want to do this, but the pipeline would need to be flexible enough to have some sort of hook | 18:53 |
| dhellmann | you might be able to build something using the existing report command | 18:54 |
| dhellmann | you would have to clone each repo, then run something like "reno report --version $v $repo" | 18:54 |
| dhellmann | well, that would give you one rst file per repo with the repo's version number | 18:55 |
| dhellmann | that wouldn't give you something easy to merge together, though | 18:55 |
| odyssey4me | dhellmann if I understand it correctly, while reno handles the heavy lifting it's sphinx that does the publishing right? | 18:56 |
| dhellmann | the new 'cache' command builds a yaml file, so you might want to use that https://review.openstack.org/289605 | 18:56 |
| palendae | odyssey4me: reno collects the ReST out of the YAML | 18:56 |
| palendae | Sphinx renders the ReST | 18:56 |
| dhellmann | odyssey4me : that's right. reno extracts the notes and formats them as rst, then sphinx converts that to html or whatever | 18:56 |
| odyssey4me | ah, that makes sense | 18:56 |
| palendae | Openstack's current sphinx invocation isn't super flexible | 18:57 |
| dhellmann | palendae : the cache command is intended to build a single yaml file for us to include in a dist where the git history won't be available downstream | 18:57 |
| dhellmann | if you wrote something to read multiple of such files, they could then be combined and rendered together | 18:58 |
| lifeless | ttx: url's ? | 18:58 |
| dhellmann | you'll need a custom release notes build job to check out all of the repos, build the yaml file, and put it somewhere | 18:58 |
| palendae | Ok, that's a great start | 18:59 |
| dhellmann | it might be better, rather than doing that when the release notes are built, to do it when a release note is added to another project and submit the results as a patch to the central release notes project | 18:59 |
| dhellmann | we're planning something similar for automating configuration reference guides | 18:59 |
| dhellmann | that's also how the requirements updates work, so there's a nice pattern there | 18:59 |
| palendae | So an automated commit is made? | 18:59 |
| dhellmann | an automated patch is submitted, but it still has to be reviewed | 18:59 |
| palendae | Oh, ok | 19:00 |
| palendae | This is already running somewhere? | 19:00 |
| palendae | Or it's planned? | 19:00 |
| dhellmann | the requirements updates are, but the config option thing is planned | 19:00 |
| odyssey4me | dhellmann that's perfectly acceptable - we're doing that manually now | 19:00 |
| palendae | Yeah, the current solution is to copy the reno file out | 19:00 |
| palendae | And manually submit 2 patches | 19:00 |
| odyssey4me | so there's just a post merge job that pulls a part of the patch out and submits it to a designated repo | 19:01 |
| dhellmann | palendae : doing it as a post-merge job also has the benefit of making the job easier to write, because you only need to check out 2 repos | 19:01 |
| palendae | If I'm following, just OSA and that role, right? | 19:01 |
| dhellmann | odyssey4me : yeah. the tricky bit will be updating an existing patch if it exists, to avoid merge conflicts, but the requirements updater already does that so you may be able to reuse the logic | 19:01 |
| dhellmann | that's right | 19:02 |
| dhellmann | you'd attach the job to the post-merge queue for the role repo and the job would check out OSA so it can submit a patch there | 19:02 |
| *** kzaitsev_mb has joined #openstack-release | 19:03 | |
| palendae | Ok, makes sense. Then the notes are duplicated, but at least in a relatively automated way | 19:03 |
| lifeless | ttx: (It's already easter here, but I'll check by a couple of times during it...) | 19:04 |
| dhellmann | palendae : as a first step, you could build tools to do that outside of CI, and then we could work on adjusting them to work in gate jobs | 19:04 |
| *** cody-somerville has quit IRC | 19:04 | |
| dhellmann | lifeless : hang on, let me find what we were talking about... | 19:04 |
| palendae | Ok, I'll see if I can take a stab at something like that | 19:05 |
| dhellmann | lifeless : ok, I have the context again: We branched openstack/requirements master to create stable/mitaka. That triggered the requirements update bot to submit a bunch of patches against other projects that also had stable/mitaka that did not match the new requirements: https://review.openstack.org/#/q/branch:stable/mitaka+topic:openstack/requirements | 19:05 |
| dhellmann | lifeless : we recommended that projects merge those changes and then submit minor version releasese from their stable branches so all projects are in sync and clear before we declare mitaka "done" | 19:06 |
| lifeless | yeah, eventual consistency. +1 | 19:06 |
| odyssey4me | palendae if you need help navigating the infra jobs, let me know - we'll need to instrument the role repositories for reno jobs which isn't hard, but I expect that'll be needed | 19:06 |
| dhellmann | lifeless : ttx wanted your thoughts on the use of a minor version update there, I think. We have used patch releases in the past, but since we don't have any new releases from master yet going with a minor version update seemed to communicate the nature of the changes better (some have new lower bound version requirements) | 19:07 |
| *** daemontool has joined #openstack-release | 19:07 | |
| *** vgridnev has joined #openstack-release | 19:07 | |
| odyssey4me | palendae I expect that pabelanger will also be happy to assist and would probably like to know about anything affecting the ansible roles, as the jobs touch his too. | 19:07 |
| lifeless | dhellmann: Perhaps I'm not clear what you mean there | 19:07 |
| dhellmann | odyssey4me , palendae : check for existing patches with the "add-reno" topic for examples of doing that. It's highly repetitive, so once you get one patch right you should be able to mostly copy those files into the other repos. | 19:07 |
| odyssey4me | dhellmann thanks for the pointers and advice, as always | 19:07 |
| lifeless | dhellmann: are you saying libraries/clients that have already done a mitaka release do a new one? | 19:07 |
| dhellmann | lifeless : yes | 19:08 |
| palendae | odyssey4me: Yeah, I've added some in the past. Mostly going to be figuring out plumbing git, and I'm gonna dig through project-config | 19:08 |
| lifeless | dhellmann: to get the new requirements out there? | 19:08 |
| dhellmann | lifeless : right. we wouldn't raise the minimum required version of the libs in the stable branch, but we would release everything that has a requirements update so they are all in sync | 19:09 |
| dhellmann | and I guess update the constraints, to go with that | 19:09 |
| lifeless | dhellmann: I don't think those additional releases are needed | 19:10 |
| lifeless | dhellmann: they're being used in concert with other libraries which have those higher mininums already | 19:10 |
| dhellmann | lifeless : http://lists.openstack.org/pipermail/openstack-dev/2016-March/090359.html | 19:10 |
| lifeless | and we know pip is just plain incapable of actually handling variation here at all - which is why we use and recommend the use of a constraints file for deployments | 19:11 |
| dhellmann | sure. the thing I was trying to avoid was in a few weeks when we start having releases from master, we didn't want anyone concerned about releasing patch releases with version updates from the stable branch | 19:11 |
| dhellmann | we sort of talked ourselves into calling those bugs in the past, but we have a chance to use the right semver rules here so I was going to take it | 19:12 |
| lifeless | k | 19:12 |
| dhellmann | so it's a bit of an extra thing, and not critical for anyone to package, but it puts us in a clean, good, well-understood state heading into newton | 19:12 |
| lifeless | I think where there are major version changes to reqs we definitely have to propogate that | 19:13 |
| lifeless | minor, as I recall, we concluded there was no formula | 19:13 |
| lifeless | library A's minor version to add a feature, is not user B's addition of a feature | 19:13 |
| dhellmann | I'm pretty sure there are no major versions but we do have cases of A needing newer minor versions of B | 19:13 |
| dhellmann | we'll have to review each release as usual, the point is really that we'd allow for something other than a patch level update in the stable branch *this time* | 19:14 |
| dhellmann | some of them are just test-requirements updates, and those don't really require a release at all, but we can sort that out when the release is actually requested | 19:14 |
| lifeless | sure | 19:18 |
| lifeless | I agree now is the time to do it | 19:18 |
| lifeless | but | 19:18 |
| lifeless | those releases won't be reachable from masters history | 19:18 |
| lifeless | so this will require tags or sem-ver: headers on master to actually get it ahead of mitaka | 19:18 |
| dhellmann | do we need to worry about that before there's something to release on master? | 19:20 |
| lifeless | (because regular commits won't outstrip a minor release) | 19:20 |
| *** cody-somerville has joined #openstack-release | 19:20 | |
| dhellmann | there's the ongoing discussion of that on the ML | 19:20 |
| lifeless | dhellmann: I don't think we want whats on PyPI to be greater than whats on master | 19:20 |
| dhellmann | hmm, yeah | 19:20 |
| lifeless | dhellmann: which might be a corner case we missed during the discussion, come to think og it | 19:20 |
| lifeless | I'm just processing that thread atm | 19:20 |
| dhellmann | yeah, I'm not sure we considered this when we talked about it | 19:21 |
| dhellmann | or if we did, we said "oh, we'll only ever release patch levels from stable branches" | 19:21 |
| dhellmann | which would still let us outpace master, of course | 19:21 |
| dhellmann | what's the concrete issue that could come up if the mitaka branch has a version > unreleased master? | 19:22 |
| *** daemontool has quit IRC | 19:25 | |
| lifeless | I just hit enter on an email about it | 19:28 |
| lifeless | but | 19:28 |
| lifeless | pip install -U foo | 19:28 |
| lifeless | where foo depends on bar | 19:28 |
| lifeless | and bar is from master git | 19:28 |
| lifeless | will 'upgrade' bar to bar from PyPI if master git is < PyPI. | 19:29 |
| dhellmann | when might that happen? | 19:32 |
| dhellmann | someone has a git url in their requirements list? | 19:33 |
| lifeless | I think its happened right now | 19:33 |
| lifeless | dhellmann: pip install . | 19:33 |
| lifeless | dhellmann: devstack uses constraints so isn't a concern | 19:33 |
| lifeless | dhellmann: folk adhocing things is where it would matter, because something they do by hand could get unwound | 19:33 |
| odyssey4me | dhellmann https://review.openstack.org/297274 verified - that's good to go whenever you're ready | 19:39 |
| odyssey4me | dhellmann https://review.openstack.org/297285 also verified - good to go whenever you're ready | 19:43 |
| dhellmann | lifeless : ok, so I'm going to agree with ttx's proposal that we hold off on approving releases until we give this more thought and see where the ML thread takes us | 19:50 |
| dhellmann | odyssey4me : ack, I'll take a look at those tomorrow when my head is clear | 19:50 |
| odyssey4me | thanks dhellmann | 19:51 |
| AJaeger | dhellmann, dims : Did you really want requirements jobs not to run on stable/mitaka? https://review.openstack.org/297225 - would like to see an ack by dhellmann on it... | 19:58 |
| dhellmann | AJaeger : yes, automatic updates should only happen on master | 19:59 |
| dhellmann | AJaeger : we may reverse that at some point, but for now it's going to be easier to keep things sane if we do it manually | 19:59 |
| AJaeger | dhellmann: ok, will +2 then and comment on your ack. | 20:00 |
| dhellmann | AJaeger : thanks for checking in, I should have +1 that when dims submitted it | 20:00 |
| AJaeger | np | 20:01 |
| *** amrith is now known as _amrith_ | 20:06 | |
| *** sdake has joined #openstack-release | 20:07 | |
| *** sdake_ has quit IRC | 20:07 | |
| *** sdake_ has joined #openstack-release | 20:10 | |
| *** sdake has quit IRC | 20:12 | |
| dhellmann | I'm going to drop off a bit early and see if I can get rid of this headache | 20:12 |
| lifeless | dhellmann: kk | 20:13 |
| lifeless | dhellmann: good luck with the headache | 20:13 |
| dims | good night dhellmann | 20:16 |
| *** kzaitsev_mb has quit IRC | 20:22 | |
| *** sridhar_ram1 is now known as sridhar_ram | 20:24 | |
| *** sdake has joined #openstack-release | 20:30 | |
| openstackgerrit | John Dickinson proposed openstack/releases: Swift 2.7.0 release https://review.openstack.org/297352 | 20:31 |
| notmyname | ttx: dhellmann: ^ swift mitaka release | 20:31 |
| *** sdake_ has quit IRC | 20:33 | |
| *** openstackgerrit has quit IRC | 20:48 | |
| *** openstackgerrit has joined #openstack-release | 20:48 | |
| *** kzaitsev_mb has joined #openstack-release | 20:50 | |
| *** doug-fis_ has joined #openstack-release | 20:57 | |
| *** doug-fish has quit IRC | 20:59 | |
| *** sdague has quit IRC | 20:59 | |
| openstackgerrit | Jim Rollenhagen proposed openstack/releases: ironic-lib 1.2.0 https://review.openstack.org/297362 | 21:01 |
| *** sdake has quit IRC | 21:01 | |
| *** kzaitsev_mb has quit IRC | 21:04 | |
| *** doug-fis_ has quit IRC | 21:05 | |
| *** sileht has quit IRC | 21:07 | |
| prometheanfire | when are upgrade docs going to be available? day of release, sooner, later? | 21:08 |
| *** sileht has joined #openstack-release | 21:08 | |
| openstackgerrit | Brant Knudson proposed openstack/releases: Keystone mitaka library update releases https://review.openstack.org/297365 | 21:10 |
| clarkb | its usually a while later as the docs team works through the process and gets it tested and written up | 21:10 |
| stevemar | bknudson: did someone ask for that ^ | 21:11 |
| bknudson | stevemar: on the mailing list | 21:11 |
| bknudson | -dev | 21:11 |
| stevemar | bknudson: oh okay, i am severely behind today | 21:12 |
| bknudson | stevemar: you need to stop working on your superman costume | 21:12 |
| stevemar | bknudson: only if you switch out of your batman one | 21:12 |
| bknudson | stevemar: http://lists.openstack.org/pipermail/openstack-dev/2016-March/090359.html | 21:13 |
| stevemar | bknudson: no keystoneclient? | 21:13 |
| *** _amrith_ is now known as amrith | 21:13 | |
| stevemar | then prepare release requests for the affected projects as soon as possible | 21:14 |
| bknudson | stevemar: https://review.openstack.org/#/q/is:merged+branch:stable/mitaka+topic:openstack/requirements -- no keystoneclient | 21:14 |
| stevemar | bknudson: i wonder why, it should have oslo.config | 21:14 |
| stevemar | and oslo.utils | 21:15 |
| bknudson | already has it - http://git.openstack.org/cgit/openstack/python-keystoneclient/tree/requirements.txt?h=stable/mitaka | 21:16 |
| stevemar | bknudson: looks like it was fine https://github.com/openstack/python-keystoneclient/commit/29e8b10750c41835b02273543c6194dae1f9fc11 | 21:16 |
| stevemar | bknudson: want to add in pycadf to that release review too? | 21:16 |
| bknudson | weird | 21:16 |
| stevemar | i just punted it through | 21:16 |
| stevemar | it should merge very soon | 21:16 |
| stevemar | its not part of integrated gate | 21:17 |
| bknudson | I'll need to wait until it merges to get the commit ID | 21:17 |
| stevemar | yeah | 21:17 |
| stevemar | bknudson: but dims and doug won't be releasing the libs any time soon | 21:17 |
| bknudson | we need to get the release requests ready ASAP and they just get to sit around?? | 21:18 |
| *** gordc has quit IRC | 21:18 | |
| stevemar | it's good to be king | 21:19 |
| *** kzaitsev_mb has joined #openstack-release | 21:21 | |
| openstackgerrit | Sean McGinnis proposed openstack/releases: Release os-brick 1.1.1 https://review.openstack.org/297371 | 21:28 |
| *** armax has quit IRC | 21:34 | |
| *** armax has joined #openstack-release | 21:34 | |
| *** devananda has quit IRC | 21:35 | |
| *** openstack has joined #openstack-release | 21:49 | |
| *** openstackgerrit has joined #openstack-release | 21:49 | |
| betherly | hi all is there anyone around who can help with the update of the requirements patch so i can re - release ironic-ui? | 21:56 |
| *** sridhar_ram has joined #openstack-release | 22:00 | |
| *** skraynev has joined #openstack-release | 22:00 | |
| openstackgerrit | Beth Elwell proposed openstack/releases: Release ironic-ui 1.2.0 https://review.openstack.org/297386 | 22:11 |
| betherly | jroll: https://review.openstack.org/#/c/297386/ | 22:11 |
| patchbot | betherly: patch 297386 - releases - Release ironic-ui 1.2.0 | 22:11 |
| jroll | betherly: +1'd with a comment | 22:13 |
| jroll | you should totally look into using reno for release notes btw :) | 22:13 |
| betherly | oh ok! thanks will do | 22:13 |
| * betherly adds that to neverending research pile | 22:13 | |
| betherly | jroll: im heading off to bed pretty soon as its 10pm here and off to surf tomorrow (YAY) if you want it to merge any sooner and it needs to be 1.1.1 would you be able to make that change for me and amend the patch so it can merge before next tuesday? | 22:15 |
| jroll | betherly: yeah, I'll wait for release team to comment, have fun surfing :) | 22:16 |
| betherly | jroll: awesome thanks! | 22:16 |
| stevemar | bknudson: pycadf should be merged soon | 22:16 |
| jroll | np | 22:17 |
| *** kzaitsev_mb has quit IRC | 22:18 | |
| *** sigmavirus24 is now known as sigmavirus24_awa | 22:31 | |
| openstackgerrit | Brant Knudson proposed openstack/releases: Keystone mitaka library update releases https://review.openstack.org/297365 | 22:40 |
| openstackgerrit | Brant Knudson proposed openstack/releases: Keystone mitaka library update releases https://review.openstack.org/297365 | 22:41 |
| *** sdake has joined #openstack-release | 22:58 | |
| *** sdake_ has joined #openstack-release | 23:17 | |
| *** sdake has quit IRC | 23:18 | |
| openstackgerrit | Jim Rollenhagen proposed openstack/releases: python-ironicclient 1.3.0 https://review.openstack.org/297452 | 23:28 |
| *** sdake has joined #openstack-release | 23:28 | |
| *** sdake_ has quit IRC | 23:28 | |
| openstackgerrit | Jim Rollenhagen proposed openstack/releases: python-ironic-inspector-client 1.6.0 https://review.openstack.org/297454 | 23:29 |
| sdake | where are the releaes notes published once you have a releasenotes job in jenkins? | 23:38 |
| jroll | sdake: docs.openstack.org/releasenotes/$project | 23:43 |
| sdake | does it upblished unreleased? | 23:47 |
| jroll | yeah, it should | 23:48 |
| jroll | e.g. http://docs.openstack.org/releasenotes/ironic/current-series.html | 23:48 |
| sdake | cool | 23:49 |
| sdake | so just need a commit to test it looks like | 23:49 |
| jroll | yep | 23:49 |
| jroll | it's a post job, so there's a delay between merge and publish | 23:49 |
| sdake | understood | 23:49 |
| *** sbezverk has quit IRC | 23:54 | |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!