*** dims has joined #openstack-release | 00:09 | |
*** sigmavirus24 is now known as sigmavirus24_awa | 00:21 | |
openstackgerrit | Merged openstack/releases: Fix pep8 errors https://review.openstack.org/266594 | 00:25 |
---|---|---|
openstackgerrit | Merged openstack/releases: copy send-announcements-to from mitaka to kilo and liberty https://review.openstack.org/266448 | 00:42 |
*** mriedem has joined #openstack-release | 01:16 | |
*** doug-fish has quit IRC | 01:18 | |
*** doug-fish has joined #openstack-release | 01:19 | |
*** doug-fish has quit IRC | 01:33 | |
openstackgerrit | Steve Martinelli proposed openstack/releases: release keystone libraries for mitaka https://review.openstack.org/266628 | 01:37 |
*** dims has quit IRC | 01:38 | |
*** LouisF has quit IRC | 01:55 | |
*** sigmavirus24_awa is now known as sigmavirus24 | 03:02 | |
*** dims has joined #openstack-release | 03:17 | |
*** mriedem has quit IRC | 03:20 | |
*** sigmavirus24 is now known as sigmavirus24_awa | 03:32 | |
*** dims has quit IRC | 03:33 | |
*** ifat_afek has joined #openstack-release | 08:36 | |
*** amotoki has joined #openstack-release | 09:10 | |
*** dtantsur|afk is now known as dtantsur | 10:02 | |
*** ifat_afek has quit IRC | 10:10 | |
openstackgerrit | Merged openstack-infra/release-tools: stop sending folks to launchpad milestones for release details https://review.openstack.org/266064 | 10:31 |
ttx | dhellmann: I'd like to get https://review.openstack.org/#/c/264754 in before it gets mergeconflicted again, so I could use your review there | 10:40 |
*** ifat_afek has joined #openstack-release | 11:06 | |
*** ifat_afek has quit IRC | 11:08 | |
*** flwang1 has quit IRC | 11:48 | |
*** flwang1 has joined #openstack-release | 11:49 | |
*** skraynev has quit IRC | 11:57 | |
sdague | lifeless: tox version is 1.9.2. tox.ini specifies min version of 1.8 | 11:59 |
*** skraynev has joined #openstack-release | 12:01 | |
*** doug-fish has joined #openstack-release | 12:01 | |
*** devananda has quit IRC | 12:08 | |
*** devananda has joined #openstack-release | 12:13 | |
*** bswartz has quit IRC | 12:19 | |
*** dims has joined #openstack-release | 12:28 | |
*** gordc has joined #openstack-release | 12:30 | |
ttx | dims: I'd like to get https://review.openstack.org/#/c/264754 in before it gets mergeconflicted again, so I could use your rereview there | 12:33 |
*** dims has quit IRC | 13:25 | |
*** dims has joined #openstack-release | 13:26 | |
*** bswartz has joined #openstack-release | 13:31 | |
ttx | dhellmann: added Daviey to the release managers groups since he will handle the last two kilo coordinated stable releases | 13:56 |
ttx | (in LP and Gerrit) | 13:56 |
dims | ttx : on my way | 13:57 |
ttx | we have been checking that the tooling will still work as far as stable/kilo is concerned | 13:57 |
dims | ttx : +2A | 13:57 |
ttx | process was only using process_bugs and upload_release anyway | 13:58 |
dhellmann | ttx: good morning | 13:58 |
dhellmann | ttx: ack on adding Daviey | 13:59 |
dhellmann | Daviey : thanks for volunteering to handle those kilo releases! | 13:59 |
Daviey | dhellmann: Thanks for having me :) | 14:00 |
dhellmann | ttx: it looks like https://review.openstack.org/#/c/264754 is failing grenade? | 14:02 |
dims | dhellmann : that's a known flaky test | 14:03 |
dims | dhellmann : that one means the vm when booted did not get an ip from dhcp | 14:03 |
dhellmann | dims : is someone working on it? I've rechecked the patch | 14:03 |
dims | dhellmann : i just figured out the logstash query a few days ago. apparently it's a well known problem with no known fixes | 14:04 |
dims | dhellmann : typically happens when a vm was booted on a specific ip and then dhcp_release is called and then a new vm is booted. the dhcp_release did not clear the dhcp lease for some reason and no one knows why this happens | 14:05 |
* dhellmann reconsiders his chosen profession | 14:06 | |
dhellmann | dims : I just noticed the note about pep8 errors, I wonder if it's a new version of pep8? | 14:07 |
dhellmann | dims : yeah, our tox.ini doesn't pin flake8 | 14:09 |
dhellmann | ttx: do you want to finish our conversation about whether to proceed with the automation? | 14:10 |
openstackgerrit | Doug Hellmann proposed openstack/releases: constrain flake8 using current global-requirements settings https://review.openstack.org/266932 | 14:12 |
dhellmann | dims : ^^ | 14:13 |
ttx | dhellmann: I was doing a consistency check on release tags and I now have a question for you | 14:18 |
dims | +2 dhellmann | 14:18 |
dhellmann | dims : thanks | 14:18 |
dhellmann | ttx: fire when ready | 14:18 |
ttx | dhellmann: we have 4 deliverables (out of 79) that are release:managed but not release:cycle* | 14:19 |
ttx | should those be considered exceptions ? Should we not have any rule there ? Should we deprecate the :managed concept anyway ? | 14:19 |
ttx | (that would be openstack-docs-theme, openstack-doc-tools, pbr and pylockfile) | 14:20 |
ttx | I think we'll soon deprecate the whole concept and handle all official projects | 14:21 |
ttx | so maybe we should not restrict :managed to :cycle* | 14:22 |
dhellmann | the managed tag sort of flipped between "ttx or dhellmann does the work" and "project follows all of the rules" | 14:22 |
ttx | yeah, a bit of overloading there | 14:23 |
dhellmann | I'd be more comfortable taking on all managed projects if we can finish the automation work. But it's not actually that much work to do one now, so maybe we can push ahead with that even if we keep doing the tagging by hand | 14:23 |
* dhellmann fidgets on the fence | 14:23 | |
ttx | the tag definition is mostly about quality | 14:24 |
ttx | not really "who will do the work" | 14:24 |
dhellmann | the other thing to consider is that right now being managed is opt-in, so we need a process for changing that to apply to all official projects | 14:24 |
dhellmann | that's what it was meant to be, but I think (at least in my head) it has several meanings | 14:25 |
ttx | ok, so I think... once we are ready to take more we'll reach out to unmanaged projects and check that they are all fine with being managed | 14:25 |
dhellmann | yeah, I'm ok with those 4 exceptions remaining independent while we continue figuring out the automation issues and communicate with the other projects | 14:26 |
*** mriedem has joined #openstack-release | 14:26 | |
ttx | if a significant proportion refuses, we'll likely keep the tag to be able to tell which | 14:26 |
dhellmann | right | 14:26 |
ttx | if only one or two refuses, we might pressure them through the tc | 14:26 |
ttx | if we have them all, we can remove the tag | 14:27 |
ttx | and/or introduce new ones to communicate release process quality if we think that matters | 14:27 |
dhellmann | yes, though we may want to do something to make clear that becoming an official project also implies having releases managed by this team | 14:27 |
ttx | sure. We would add that to the new team process | 14:28 |
dhellmann | we could do that at the same time as removing the tag | 14:28 |
ttx | once we have all the existing ones signed up | 14:28 |
dhellmann | right | 14:28 |
dhellmann | in the time since yesterday's infra meeting, I think I've come to terms with the idea of just continuing to run the release scripts by hand | 14:28 |
dhellmann | with the new email automation, it's just one command, so the only "hard" part is hanging around and waiting for the release requests to merge | 14:29 |
dhellmann | which just means I'll stop doing them late in the afternoon when zuul is really busy | 14:29 |
dhellmann | ttx, dims: what do you think? | 14:29 |
ttx | Lets think a bit more about it | 14:29 |
ttx | dhellmann: what is the part that is painful to automate exactly ? | 14:30 |
dims | dhellmann : ack one command does make it easy | 14:30 |
ttx | signed tagging ? | 14:30 |
dhellmann | the hard part right now is actually getting the tools we've built into the images | 14:30 |
dhellmann | the CI images | 14:30 |
dhellmann | we can't install things from pip, so we need system packages for all our deps | 14:31 |
dims | dhellmann : so yes, i'll help with the day to day running commands / waiting etc :) | 14:31 |
dhellmann | I haven't checked the others, but I'd be surprised if there was a .deb for reno yet | 14:31 |
ttx | dhellmann: sure, but which tools | 14:31 |
dhellmann | and reno is now a dep for the announce script | 14:31 |
ttx | let's take the process in steps | 14:31 |
dhellmann | ttx: well, all of them, right? we wanted the whole process handled by the CI servers | 14:31 |
ttx | we start with the releases post-merge process | 14:32 |
ttx | which is supposed to generate a tag | 14:32 |
dhellmann | release_from_yaml.sh runs some python code to run release.sh to generate the tag | 14:32 |
dims | dhellmann : i can ask Thomas Goirand (zigo) for help with .deb | 14:32 |
dhellmann | I think release_from_yaml.sh just needs pyyaml, so we should be ok there | 14:32 |
dhellmann | dims : ack, we'd need to get it into the canonical repos somehow, too | 14:33 |
ttx | IIUC the painful part is that some of the jobs need access to secrets | 14:33 |
dims | dhellmann : ah i see | 14:33 |
ttx | secrets are used for signing | 14:33 |
ttx | we sign two things: the tags and the tarballs | 14:33 |
dhellmann | there are also launchpad credentials to manage, but yeah | 14:34 |
ttx | we already established that tarball signing could still be automated | 14:34 |
ttx | since it's not really our scripts | 14:34 |
ttx | that leaves two parts: tagging and announcing | 14:34 |
dhellmann | right, that's happening as part of a related blueprint that infra is doing anyway | 14:34 |
ttx | tagging happens before, announcing after the tarball signing part | 14:34 |
ttx | at least that would be best that way | 14:35 |
dhellmann | yes | 14:35 |
ttx | we could trigger the announce job from the tag post-job | 14:36 |
dhellmann | that's not the order it happens now, but we wanted to fix that | 14:36 |
ttx | that's a bit out of sequence since there is a race with tarball publication | 14:36 |
* dhellmann looks at what args the announce needs | 14:36 | |
dhellmann | announce.sh takes the repository directory and the version number | 14:37 |
ttx | we could also have a specific jenkins job that reacts to the tarball publication but would still run on an unprividleged node | 14:37 |
ttx | so options for announcements are: | 14:37 |
ttx | 1/ manual announce, watching manually for tarball to be published by automation | 14:38 |
ttx | 2/ automatic announce even before the tarball is published | 14:38 |
ttx | 3/ automatic announce from the tarball publication job, on the privileged node (with our tooling in project-config) | 14:39 |
ttx | 4/ automatic announce from some separate unprivileged job that watches tarball creation | 14:39 |
ttx | we ruled 3 out as not agile enough | 14:39 |
ttx | 2 introduces a race and we can't post the characteristics of the tarball in the announcement | 14:39 |
dhellmann | I'm not sure 2 would work, based on the way that script works today -- it needs the tag in place to compute the differences | 14:40 |
ttx | that leaves 1 and 4 | 14:40 |
ttx | (-for the announce part) | 14:40 |
ttx | Now let's take the tagging problem | 14:40 |
ttx | we want tagging to happen automatically when the openstack/releases change merges | 14:40 |
ttx | but we want signed tags too | 14:41 |
dhellmann | right | 14:41 |
ttx | signed tags mean priviledged nodes, or precomputed signatures | 14:41 |
ttx | i.e. include the signature in the request and somehow just apply it within the tag | 14:42 |
dhellmann | assuming infra works out key management, the tagging part isn't a problem from the perspective of moving the tools into the project-config repo | 14:42 |
ttx | we'd basically work around gerrit not supporting tag reviews | 14:42 |
dhellmann | we might need some changes, but they shouldn't be major | 14:42 |
ttx | dhellmann: why? | 14:42 |
ttx | because you don't expect that many changes there ? | 14:42 |
dhellmann | right: we need to move the things that are console scripts to regular python scripts that can be run without an installation step, and we need the bash scripts to not build virtualenvs but assume all the needed packages are installed system-wide | 14:43 |
ttx | dhellmann: hmm, ok. | 14:43 |
ttx | so for the tagging, solutions are: | 14:44 |
ttx | 1/ manual tagging once the releases change merges | 14:44 |
ttx | 3/ automatic tagging from a privileged node | 14:44 |
ttx | 5/ some precomputation of the tag being present in the change and applied on a unprivileged node | 14:45 |
ttx | this 3 being less painful that the other 3 | 14:45 |
ttx | than* | 14:45 |
dhellmann | at the summit we rejected 5 as unnecessarily complex, given that we need secrets for signing tarballs anyway | 14:45 |
ttx | lets call the first set of solutions A (like announce) and the second set T (like tagging) | 14:46 |
dhellmann | and right, the tools for introducing the tags are less complex so I think moving those won't be as bad as moving the announce scripts | 14:46 |
ttx | Wa are at T1 A1 right now | 14:46 |
ttx | We should shoot for T3 first | 14:46 |
ttx | (and still do A1) | 14:47 |
ttx | wait for tarball signature/upload to be completed on infra side | 14:47 |
ttx | then consider moving from A1 to either A3 or A3 | 14:47 |
openstackgerrit | Graham Hayes proposed openstack/releases: Release designate 1.0.1 https://review.openstack.org/266952 | 14:47 |
ttx | A4 or A3 I mean | 14:47 |
dhellmann | ok, that gives us an incremental improvement, I suppose | 14:48 |
ttx | so... T1A1 -> T3A1 -> maybe T3A3 or T3A5 | 14:48 |
dhellmann | I wonder if we could have A4 using a job on the releases repo, so we would know what version info we're working with | 14:48 |
ttx | it's also easier to parallelize with the work on tarball signature | 14:48 |
ttx | err T3A4 you are right | 14:49 |
ttx | A4 would need to be triggered by either the tarball upload job completion, or detecting the presence of the file | 14:50 |
dhellmann | right, I was thinking it would be triggered when the release change merged and then it would wait for the tarball to be present before continuing | 14:50 |
ttx | oh. | 14:50 |
dhellmann | that way it has the patch from the releases repo to tell it what was released | 14:50 |
ttx | sounds a bit brittle, we would have to check that all steps completed. | 14:51 |
dhellmann | it might need to watch for a bunch of releases, like when dims does the oslo weeklys | 14:51 |
ttx | But then we need to keep an eye on release job failures anyway | 14:51 |
dhellmann | yeah, that's a good point, we need some good way to track the job outcomes to make sure none fail | 14:51 |
dhellmann | otoh, we could write a job like I describe for A4 today | 14:52 |
ttx | anyway, i would keep on doing A1 until T is automated | 14:52 |
dhellmann | it can just run in a tox environment, so we wouldn't need much special tooling in infra at all | 14:52 |
ttx | right | 14:53 |
ttx | the trick is to separate what needs to run on the privileged node (not that much) and what doesn't | 14:53 |
dhellmann | ok, I'll go back and look at the infra spec and see if it needs to be changed to reflect this change in plan | 14:53 |
ttx | ok | 14:54 |
dhellmann | I'll also see if I can build a release_from_yaml.sh with fewer dependencies | 14:54 |
ttx | and I enter a meeting tunnel | 14:54 |
ttx | be back tomorrow :) | 14:54 |
dhellmann | have fun? :-) | 14:54 |
openstackgerrit | Graham Hayes proposed openstack/releases: Release designate 2.0.0.0b2 https://review.openstack.org/266963 | 15:05 |
openstackgerrit | Davanum Srinivas (dims) proposed openstack/releases: Oslo Releases for week of Jan 11, 2016 https://review.openstack.org/265673 | 15:06 |
*** devananda has quit IRC | 15:59 | |
*** devananda has joined #openstack-release | 16:02 | |
*** sigmavirus24_awa is now known as sigmavirus24 | 16:07 | |
mugsie | ttx: you around? - getting an error for my stable/liberty release patch, and I can't see what I am missing | 16:09 |
mugsie | http://logs.openstack.org/52/266952/1/check/gate-releases-tox-validate/b35f77d/console.html are the logs | 16:09 |
mugsie | review: https://review.openstack.org/266952 | 16:09 |
ttx | mugsie: looking | 16:10 |
mugsie | ty | 16:10 |
ttx | mugsie: it's missing a send-announcements-to field. see http://lists.openstack.org/pipermail/openstack-dev/2016-January/083749.html | 16:12 |
ttx | see for example http://git.openstack.org/cgit/openstack/releases/tree/deliverables/mitaka/designate.yaml | 16:13 |
mugsie | yeah, saw the field, missed the email, and assumed it was an optional field | 16:13 |
mugsie | thanks, will update it | 16:13 |
ttx | I'm surprised that it's not optional, since we have a default value | 16:14 |
ttx | but yeah, that is probably what the error is about | 16:14 |
openstackgerrit | Steve Martinelli proposed openstack/releases: release keystone libraries for mitaka https://review.openstack.org/266628 | 16:14 |
stevemar | dhellmann: dims ^ requirements patch is up as well, and it should be passing jenkins | 16:15 |
openstackgerrit | Graham Hayes proposed openstack/releases: Release designate 1.0.1 https://review.openstack.org/266952 | 16:16 |
SergeyLukjanov | ttx, dhellmann, hi, we've found out that we created a new repo (sahara-scenario) incorrectly and we need to have actually sahara-tests repo with git history, do you think it'll be okay to request new repo and move old one to attic enventually? | 16:17 |
SergeyLukjanov | (shame me, was travelling too much last month and missed it) | 16:17 |
*** dims_ has joined #openstack-release | 16:18 | |
ttx | SergeyLukjanov: you should first check with infra on the feasibility... it doesn't trigger that many alarms from where I stand | 16:20 |
SergeyLukjanov | yeah, probably, it'll be easier to rename repo and re-push with --force | 16:21 |
SergeyLukjanov | ttx, thx | 16:21 |
*** dims has quit IRC | 16:21 | |
dims_ | stevemar : will look in a little bit, gotta run an errand | 16:24 |
stevemar | dims_: cool cool | 16:28 |
*** dtantsur is now known as dtantsur|afk | 16:54 | |
dims_ | stevemar : +2 hoping dhellmann will have time to cut it else i can do it later this evening/night | 16:56 |
stevemar | dims_: cool cool | 16:57 |
stevemar | dims_: thanks for looking at it :) | 16:57 |
dhellmann | stevemar, dims_ : I'm in a meeting, but will look at it right afterwards | 16:57 |
stevemar | dhellmann: dims_ y'all are the best :) | 16:57 |
dims_ | :) | 16:58 |
*** amotoki has quit IRC | 17:11 | |
dhellmann | dims_ , stevemar : starting those releases now | 17:12 |
*** fesp has joined #openstack-release | 17:32 | |
*** fesp has quit IRC | 17:34 | |
openstackgerrit | Merged openstack/releases: constrain flake8 using current global-requirements settings https://review.openstack.org/266932 | 17:43 |
openstackgerrit | Merged openstack/releases: Release designate 1.0.1 https://review.openstack.org/266952 | 17:52 |
*** dougwig has quit IRC | 18:08 | |
*** lane_kong has quit IRC | 18:08 | |
dhellmann | mugsie: hey, the release announcement for designate failed because there's no README.rst file: http://paste.openstack.org/show/483803/ | 18:21 |
mugsie | dhellmann: I thought we had a readme ... | 18:22 |
dhellmann | you have a README.md | 18:22 |
mugsie | oh. its in markdown | 18:22 |
mugsie | OK, so we should move that to rst then? | 18:22 |
dhellmann | yes, please | 18:23 |
dhellmann | the announce script looks for some links in a specific format, too, so check some of the other projects to see how they do things like link to the bug tracker | 18:23 |
*** LouisF has joined #openstack-release | 18:24 | |
mugsie | ack | 18:24 |
*** dougwig has joined #openstack-release | 18:25 | |
sdague | does anyone know the minimum tox required for constraints to work? | 18:25 |
mugsie | dhellmann: I just checked nova and cinder and they are both completely different | 18:26 |
dhellmann | mugsie: have a look at one of the oslo libs, I made sure those were all right | 18:26 |
*** lane_kong has joined #openstack-release | 18:26 | |
dhellmann | mugsie : you want a block like this somewhere: http://paste.openstack.org/show/483805/ | 18:27 |
dhellmann | sdague : tox shouldn't matter, but pip will | 18:27 |
mugsie | ah, ok. will add that now | 18:27 |
sdague | dhellmann: well I have pip 7.1.2 | 18:28 |
sdague | but tox -e py27-constraints doesn't work | 18:28 |
dhellmann | sdague : what's not working, I can try to reproduce | 18:28 |
sdague | git clone nova | 18:28 |
sdague | tox -e py27-constraints | 18:28 |
sdague | explode | 18:28 |
dhellmann | ok, running that now | 18:29 |
sdague | http://paste.openstack.org/show/483806/ | 18:29 |
lifeless | sdague: I suspect you've found a limitation in older tox that we didn't test far enough back to find | 18:29 |
lifeless | sdague: sdague did you try a current tox? | 18:29 |
sdague | right, so that was the question. I have 1.9.1, the tox.ini says min tox is 1.8 | 18:29 |
sdague | I have not yet, I was asking the question what the minimum is | 18:30 |
lifeless | sdague: given when we developed it, perhaps try a 2.x ? | 18:30 |
lifeless | sdague: I don't know the actual minimum, except via trial and error | 18:30 |
*** dims_ has quit IRC | 18:30 | |
sdague | ok, I guess I assumed that would have been done as part of the effort, because local run is basically the blocker for nova now | 18:31 |
lifeless | sdague: I don't think we actually had 'test with tox older than default' as an identified work item | 18:31 |
lifeless | sdague: sorry | 18:31 |
*** dims has joined #openstack-release | 18:32 | |
sdague | lifeless: it seems ironic that the effort working on constraints wouldn't consider the [tox] | 18:32 |
sdague | minversion = 1.8 | 18:32 |
sdague | bit :) | 18:32 |
lifeless | sdague: a little | 18:34 |
lifeless | sdague: OTOH it is one of the things the jenkins slaves standardise on updating upfront | 18:34 |
sdague | right, but this is a user callable target | 18:35 |
sdague | this isn't just about the gate, it's about local dev env acting like the gate | 18:35 |
lifeless | sdague: I've already admitted error, I feel like you're flogging the point | 18:35 |
sdague | 2.0 seems to have gotten past the old error | 18:35 |
lifeless | great | 18:35 |
LouisF | i want to get networking-sfc released to mitaka, do i need to add networking-sfc.yaml to deliverables/mitaka? | 18:38 |
lifeless | LouisF: yes | 18:38 |
dhellmann | LouisF : there are some instructions here: http://docs.openstack.org/releases/instructions.html#requesting-a-release have you had a chance to look at those? | 18:39 |
LouisF | yes looked at those, just verifying that is correct procedure | 18:39 |
dhellmann | ok, yes, assuming networking-sfc is a new deliverable and not related to something that already exists, you would create a new file | 18:40 |
LouisF | the git commit hash used in the yaml is obtained from a final commit to networking-sfc? | 18:41 |
dhellmann | LouisF : yes | 18:42 |
dhellmann | LouisF : make sure to use the merge commit sha, if there is a merge commit | 18:42 |
LouisF | ok thanks, when should this completed to be in M2? | 18:43 |
*** ig0r_ has joined #openstack-release | 18:43 | |
dhellmann | LouisF : the schedule is at http://docs.openstack.org/releases/schedules/mitaka.html -- the M2 deadline is Jan 22 | 18:43 |
LouisF | dhellmann: thanks | 18:44 |
dhellmann | mugsie : looking at the beta release for designate, I'm a little confused. According to http://governance.openstack.org/reference/projects/designate.html the dashboard is a separate deliverable from designate itself, and has different release tags. Is that right? | 18:46 |
dhellmann | LouisF : np, let me know if you run into any other questions | 18:46 |
mugsie | dhellmann: it is separate now? | 18:46 |
dhellmann | mugsie : yet the release request includes both items together | 18:46 |
dhellmann | mugsie : do designate and designate-dashboard share a launchpad page? | 18:47 |
dhellmann | and more importantly, bug tracker | 18:47 |
mugsie | yes | 18:47 |
dhellmann | ok | 18:47 |
mugsie | they were combined before | 18:47 |
mugsie | (we did m1 as a combined deliverable) | 18:48 |
dhellmann | so that's ok for the release repo then | 18:48 |
dhellmann | I'm not sure about the tags, though. it would be better if they reflected the fact that they use the same release model | 18:48 |
mugsie | yeah | 18:48 |
dhellmann | yeah, I'm just confused because the 2 sets of instructions don't say the same thing -- governance thinks they are different, release thinks they are together | 18:48 |
mugsie | As far as I was concered - they were the same thing | 18:49 |
dhellmann | so all of the release tags apply the same? cycle-with-milestones and has-stable-branches? | 18:49 |
dhellmann | mugsie : maybe you can submit the governance repo patch to straighten that out? | 18:50 |
mugsie | yeah. thats how we have been operating. I will make a change to the projects.yaml file later on today | 18:50 |
dhellmann | mugsie : ok, great, thanks | 18:51 |
mugsie | np - thanks for letting me know | 18:51 |
dhellmann | mugsie : ok, I commented on the release request for the rest of the release team to know we're waiting | 18:52 |
LouisF | dhellmann: this is a new deliverable - subproject of neutron | 18:53 |
lifeless | sdague: did it finish successfully, or bail ? | 18:53 |
dhellmann | LouisF : ok, that's what I thought based on the name, I just wasn't sure if it had ever been released before | 18:53 |
sdague | it looks successful | 18:53 |
LouisF | dhellmann: ok so this is correct procedure? | 18:53 |
lifeless | sdague: ok, so do we care about bisecting for the min tox, or just say 2.0 is the min everywhere? | 18:54 |
dhellmann | LouisF : yes, please create a new file following those instructions | 18:54 |
sdague | I had to pull a second tree, my primary tree is in the middle of a chunk of work, so tests aren't passing there for other reasons | 18:54 |
LouisF | dhellmann: thanks again | 18:54 |
sdague | 2.0 seems fine, I was on a 1.9.2 already, I don't even know how many bumps there are in between | 18:54 |
lifeless | sdague: ok, i'll submit a change to cookie-cutter | 18:55 |
lifeless | sdague: I presume you'll put up a nova patch? Or would you like me to? | 18:56 |
sdague | yeh, I've got the nova patch up already | 18:57 |
sdague | the nova team is also leaning heavily on just using this for everything instead of making it a dev option to get wrong | 18:57 |
lifeless | cool | 18:58 |
sdague | I just proposed https://review.openstack.org/#/q/status:open+project:openstack/nova+branch:master+topic:tox_constraints to get us over there | 19:00 |
*** mriedem is now known as mriedem_meeting | 19:02 | |
lifeless | sdague: yeah, so I'm concerned about that without zuul/jjb glue being done | 19:03 |
lifeless | sdague: I don't know if you saw my replies to your question about doing htat | 19:03 |
sdague | lifeless: right, I'm not sure I understand why that would be different there | 19:03 |
sdague | I guess there is a set of details that I am missing | 19:03 |
lifeless | sdague: have a look at jenkins/scripts/run-tox.sh | 19:05 |
lifeless | sdague: and, let me find the othe rbit | 19:05 |
lifeless | sdague: ah yes - zuul-git-prep-upper-constraints | 19:05 |
lifeless | sdague: the constraints file is from a different repo | 19:06 |
sdague | run-tox.sh always exports the upper constraints file | 19:06 |
lifeless | sdague: yup, which won't exist unless you've run zuul-git-prep-upper-constraints in the job definition | 19:06 |
lifeless | sdague: so ignoring the potential dev confusion - which honestly I don't care either way on | 19:07 |
sdague | ok, so where do I get agreement for adding this to the other jobs? | 19:07 |
lifeless | sdague: making sure it actually works is the bit that concerns me | 19:07 |
sdague | right, sure, that's fair | 19:08 |
sdague | some set of testing will happen with these patches | 19:08 |
lifeless | sdague: infra I suspect are the best human proxies for figuring it out | 19:08 |
mugsie | dhellmann: proposed https://review.openstack.org/267105 | 19:09 |
sdague | fungi: is pinging there, I'll context switch to it | 19:09 |
lifeless | sdague: yah | 19:09 |
dhellmann | mugsie : thanks | 19:14 |
fungi | sdague: yeah, sorry didn't realize it was being discussed here too | 19:17 |
fungi | unfortunately i don't have bandwidth to discuss it at the moment | 19:18 |
*** mriedem_meeting is now known as mriedem | 19:52 | |
*** dims has quit IRC | 20:18 | |
*** dims has joined #openstack-release | 20:18 | |
openstackgerrit | Doug Hellmann proposed openstack-infra/release-tools: refactor list_deliverable_changes.py prior to move to project-config https://review.openstack.org/267161 | 20:49 |
openstackgerrit | Doug Hellmann proposed openstack-infra/release-tools: refactor script for commenting on launchpad bugs prior to move to project-config https://review.openstack.org/267162 | 20:49 |
openstackgerrit | Doug Hellmann proposed openstack-infra/release-tools: default to sending mail directly via lists.openstack.org https://review.openstack.org/267166 | 20:55 |
dhellmann | dims : it would be good to test a release using these changes ^^ (though maybe not the oslo megapack release) | 20:57 |
openstackgerrit | gordon chung proposed openstack/releases: ceilometerclient 2.2.1 https://review.openstack.org/267168 | 20:57 |
dims | dhellmann : right, will review those today | 20:57 |
dhellmann | dims : great, I'd rather not merge them until we've done a practical test | 20:58 |
dims | dhellmann : good point | 20:58 |
*** ig0r_ has quit IRC | 21:14 | |
*** ig0r_ has joined #openstack-release | 21:23 | |
*** ig0r_ has quit IRC | 21:31 | |
*** ig0r_ has joined #openstack-release | 21:45 | |
*** ig0r_ has quit IRC | 21:45 | |
openstackgerrit | Monty Taylor proposed openstack/releases: Release os-client-config 0.14.1 https://review.openstack.org/267196 | 22:09 |
*** sigmavirus24 is now known as sigmavirus24_awa | 22:55 | |
*** mriedem is now known as mriedem_away | 23:00 | |
*** amotoki has joined #openstack-release | 23:28 | |
*** gordc has quit IRC | 23:29 | |
*** topol has quit IRC | 23:34 | |
*** LouisF has quit IRC | 23:36 | |
*** topol_ has joined #openstack-release | 23:37 | |
*** topol_ is now known as Guest57135 | 23:37 | |
*** amotoki has quit IRC | 23:52 | |
*** amotoki has joined #openstack-release | 23:52 | |
*** amotoki has quit IRC | 23:53 |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!