Tuesday, 2022-04-19

Clark[m]fungi: looks like the storyboard fixup change passes00:10
opendevreviewIan Wienand proposed opendev/lodgeit master: new paste: add maxlength to input box  https://review.opendev.org/c/opendev/lodgeit/+/83841800:29
ianwi guess our daily builds have been isolated from git updates until we merge a nodepool change that rebuilds the container?  looks like the last was apr 1300:40
ianweven then, i wonder what layer git comes in at; is it already in the published images?  maybe we have to rebuild our python-builder environment too?00:40
Clark[m]Last I checked debian hadn't updated01:01
Clark[m]Which means we're isolated01:02
ianwoh, well that could explain it too :)  i'd assumed they'd pulled the same update as ubuntu01:05
Clark[m]It's possible they have since I checked on Friday or that they do tomorrow01:09
Clark[m]But I decided to punt on that since debian packages weren't updated01:10
Clark[m]I think that is the right changeling for bullseye01:10
Clark[m]Heh auto correct changelog to changeling01:11
ianwyeah https://security-tracker.debian.org/tracker/CVE-2022-2476501:11
opendevreviewIan Wienand proposed opendev/lodgeit master: new paste: check input length  https://review.opendev.org/c/opendev/lodgeit/+/83842101:30
opendevreviewIan Wienand proposed opendev/lodgeit master: new paste: check input length  https://review.opendev.org/c/opendev/lodgeit/+/83842102:52
*** pojadhav is now known as pojadhav|ruck03:59
*** chandankumar is now known as chkumar|ruck04:47
*** chkumar|ruck is now known as chandankumar04:51
*** ysandeep|out is now known as ysandeep05:49
dpawlikgmann: hey, catch me when you are online06:14
*** jpena|off is now known as jpena07:01
*** ysandeep is now known as ysandeep|lunch07:29
*** pojadhav|ruck is now known as pojadhav|lunch07:35
*** arxcruz|out is now known as arxcruz08:11
opendevreviewAlfredo Moralejo proposed zuul/zuul-jobs master: Update gpg key file for extras-common in CS9  https://review.opendev.org/c/zuul/zuul-jobs/+/83845008:19
*** pojadhav|lunch is now known as pojadhav|ruck08:30
*** ysandeep|lunch is now known as ysandeep09:50
*** icey_ is now known as icey09:53
*** dviroel|out is now known as dviroel11:31
*** ysandeep is now known as ysandeep|afk11:41
hrwcan someone check https://mirror.regionone.osuosl.opendev.org/wheel/centos-8-aarch64/grpcio/ dir? 2 files, 0 bytes each == jobs fail12:30
hrwhttps://mirror.iad3.inmotion.opendev.org/wheel/centos-8-aarch64/grpcio/ has 0 bytes too12:33
hrwcan those files be removed?12:33
fungiagreed, those files are 0 bytes in /afs/openstack.org/mirror/wheel/centos-8-aarch64/g/grpcio/12:47
fungiwith creation dates of august 13/14 of 202112:47
fungineed to dig out the docs on deleting files from afs12:48
hrwand we did not built that old branch a while as newer versions are wheeled by upstream12:51
*** ysandeep|afk is now known as ysandeep12:53
fungi#status log Manually deleted /afs/openstack.org/mirror/wheel/centos-8-aarch64/g/grpcio as it contained only corrupt (zero-byte) packages12:55
opendevstatusfungi: finished logging12:55
fungithanks for pointing it out!12:55
hrwI did not looked for other 0 ones - those were where builds failed12:56
*** rcastillo|ruck is now known as rcastillo|rover13:47
*** dviroel is now known as dviroel|afk14:24
*** dviroel|afk is now known as dviroel|mtg14:39
*** ysandeep- is now known as ysandeep|out14:45
*** pojadhav|ruck is now known as pojadhav|out14:49
clarkbhrw: as a side note the same filesystem servers all of the mirrors so you should see consistent info across them. If you don't that would be a bug14:50
*** dviroel|mtg is now known as dviroel14:54
clarkbdpawlik: gmann: re https://review.opendev.org/c/opendev/system-config/+/838324 and https://review.opendev.org/c/opendev/base-jobs/+/837619 can people use the new system today? If so can we announce it and start to shut down the old system? I really don't want to keep it around much longer since it is impacting jobs now15:02
clarkband we're past the agreed upon transition point (yoga release)15:02
gmanndpawlik: clarkb yeah, so i am waiting for the redirect question on this https://review.opendev.org/c/openstack/governance-sigs/+/83583815:03
clarkbkopecmartin: when your meeting is done did you want to review https://review.opendev.org/c/opendev/system-config/+/838344 which updates our refstack deployment images to python3.9?15:03
gmannI used the new dashbaord and I think it is good to start using. dpawlik can give more updates if anything we need to waity15:03
clarkbgmann: dpawlik  note we wouldn't host an http redirect. We can update dns to point logstash.openstack.org to a different name/ip though15:04
kopecmartinclarkb: sure15:04
clarkba redirect implies an http server at the old name doing the redirect and we're saying we want to delete the host currently at that name15:04
clarkbgmann: dpawlik: but also updating the dns name can happen after the fact. I don't think that hsould hold anything up?15:06
fungiwhat i considered was cnaming the old name to the new dashboard, or to static.opendev.org and adding it as a redirect alias to either the new dashboard or the sig page with the conneciton info, but we could also just announce that the old name is no longer in use and tell people where to start going15:06
*** dviroel is now known as dviroel|lunch15:06
clarkbright the problem with just a CNAME or dns update is likely going to be ssl15:07
clarkbbut I don't think opendev should host a redirect either15:07
dpawlikIMHO the new Elasticsearch is ready to use15:07
clarkbbut also I really don't think that should hold up any change15:07
clarkbsince you ahve to know how to login to the new thing anyway using a new url isn't that big of a deal15:07
clarkbif you do get redirected you'll just get stuck on logging in15:07
fungithe old dashboard is broken since weeks at least, so i have doubts there are that many people who will be surprised if it just disappears15:07
clarkbdpawlik: that is greaet to hear :)15:08
clarkbdpawlik: maybe you can +1 those two changes I linked indicating you are ready for users to use the opensearch system now? that would be good record keeping15:08
gmannyeah, let me put the plan in TC meeting on thursday and we can discuss about communicating the new dashboard to community. 15:08
fungibut communication planning shouldn't delay taking down an already broken service, right?15:08
gmannI also think we are good to go now. agree on redirect is not something blocker 15:08
funginobody is able to use the old service now, and we have no plans to spend time fixing it15:09
dpawlikLast week I was trying to include the the openstack user into the private tenant, so user no matter where it connect to the opensearch dashboard will not force to change the tenant to global15:09
dpawlikclarkb: ok15:09
clarkbdpawlik: https://review.opendev.org/c/opendev/system-config/+/838324 and https://review.opendev.org/c/opendev/base-jobs/+/837619 are the two changes I referred to15:10
gmanni know things are broken now but waiting for communicating the new things does not harm which is just 2 days after or so15:11
gmanndpawlik: fungi so this is good to go right? I will +W https://review.opendev.org/c/openstack/governance-sigs/+/83583815:12
dpawlikgmann: from my side yes15:13
gmannok, approving that15:14
gmanndpawlik: clarkb fungi added it in TC thursday meeting agenda https://wiki.openstack.org/wiki/Meetings/TechnicalCommittee#Agenda_Suggestions15:14
fungigmann: yes, that's what my code review +1 was meant to indicate. as tact sig chair, i'm in favor of approving it15:16
dpawlikcool. Glad that the topica is moving forward 15:16
fungimmm, tapioca15:16
clarkbapparently git-review's new get_git_version doesn't work on OSX because apple patches git to print different version info15:36
clarkbI'll work on a fix between meetings15:36
fungioh fun :/15:37
fungii'm close to having the release doc builds i think, will push up some changes monentarily15:39
opendevreviewClark Boylan proposed opendev/git-review master: Fix get_git_version on OS X  https://review.opendev.org/c/opendev/git-review/+/83852315:47
clarkbsomething like ^ maybe. I'm trying to avoid having to write a super complicated parser that ends up breaking because it is complicated15:47
clarkbI suspect this is a good balance of simple and close enough. SHould work as long as we don't start prefixing the version with extra platform specific info too15:48
*** marios is now known as marios|out15:48
clarkband now breakfast before the next meeting15:49
hrwclarkb: wanted to check in case one of mirrors went out of space or weird15:49
opendevreviewJeremy Stanley proposed opendev/base-jobs master: Add a opendev-publish-unversioned-tox-docs job  https://review.opendev.org/c/opendev/base-jobs/+/83852515:53
opendevreviewJeremy Stanley proposed opendev/base-jobs master: Rename opendev-zuul-docs to opendev-versioned-docs  https://review.opendev.org/c/opendev/base-jobs/+/83852615:53
opendevreviewJeremy Stanley proposed opendev/git-review master: Run opendev-publish-tox-docs on release  https://review.opendev.org/c/opendev/git-review/+/83834115:54
opendevreviewJeremy Stanley proposed opendev/bindep master: Publish documentation at every release  https://review.opendev.org/c/opendev/bindep/+/83834015:55
opendevreviewJeremy Stanley proposed opendev/git-review master: Publish documentation at every release  https://review.opendev.org/c/opendev/git-review/+/83834115:56
mnaserclarkb: you can kill those images, i don't think we've used them for a little bit16:06
*** jpena is now known as jpena|off16:08
opendevreviewJeremy Stanley proposed opendev/base-jobs master: Add a opendev-publish-unversioned-tox-docs job  https://review.opendev.org/c/opendev/base-jobs/+/83852516:13
opendevreviewJeremy Stanley proposed opendev/base-jobs master: Rename opendev-zuul-docs to opendev-versioned-docs  https://review.opendev.org/c/opendev/base-jobs/+/83852616:13
*** dviroel|lunch is now known as dviroel16:14
clarkbmnaser: thank you for confirming. This will help us trim down our image list quite a bit before adding python3.1016:15
clarkbfungi: shoudl we approve https://review.opendev.org/c/opendev/git-review/+/838523 and do a quick release?16:17
clarkbkopecmartin: re https://review.opendev.org/c/opendev/system-config/+/838344 I checked and refstack runs tests on python3.9 now so I don't really expect any issues. But definitely point out if I've missed something16:18
fungiclarkb: yeah, though i'd like to try to get 838341 in too before we tag16:18
fungisince that will allow us to quickly test if it's a viable approach16:19
clarkbgood point16:21
clarkblet me pull those up now16:21
clarkbfungi: looks like we set the branch and tag values to always be latest? that makes sense to me16:24
clarkbI've +2'd the two changes necessary for git-review to get that done. I guess we can alnd those then check it works and if so work to land the others16:24
fungii left a comment with a caveat on 838525, i don't think it's a blocker but food for thought16:25
clarkbya I think we can live with that16:25
clarkbit is a rare case, but the case of witing on doc updates is not rare16:25
fungiand i think it's solveable, but will need some edits to the ansible16:26
fungialso that job is probably not applicable for projects which tag releases on multiple branches, but again that's rare for opendev's own tool/utility projects16:27
clarkbI'm going to delete the held nodes that I set up to debug git problems last week16:27
clarkblooks like the gear tests failed on the dockerfile update too16:33
clarkbit was a job timeout so I'm rechecking to see if it is consistent but will try and dig into that a bit like I did with storyboard16:33
clarkbhttps://review.opendev.org/c/opendev/storyboard/+/838409 is the storyboard fix. Nothing too crazy in there and I tried to explain each chagne inthe commit message16:34
opendevreviewMerged opendev/git-review master: Fix get_git_version on OS X  https://review.opendev.org/c/opendev/git-review/+/83852316:36
opendevreviewMerged opendev/base-jobs master: Add a opendev-publish-unversioned-tox-docs job  https://review.opendev.org/c/opendev/base-jobs/+/83852516:40
fungiclarkb: should we be considering retiring gear now that we don't use it?16:48
clarkbfungi: its not a bad idea, though we may want to do that independent of this effort since I think some people have been using it16:49
clarkbbasically take some time to communicate it? Though maybe no one is using it and I'm mistaken16:49
clarkbit does look like gear testing doesn't work with python3.9 without timing out. The other versions take just a few minutes to run16:50
clarkbin theory it did work on python3.9 though as testing would've had to pass to add the job16:50
fungiwell, also if people are still using it and need new versions/fixes in it, we should perhaps encourage them to take it over16:50
clarkbI cannot reproduce this timeout issue locally on tumbleweed with python3.9 either16:52
fungionce 838341 merges i'll prepare to tag a git-review 2.3.116:52
fungiclarkb: oh, hrm, 838523 did not include a release note. i expect we want one?16:53
clarkboh ya. I always forget these small tools use reno16:54
fungii'm about to push one now16:54
clarkbok I can reproduce the issue if I update setuptools in my tox venv16:54
clarkbthe issue is running python setup.py testr is broken with current setuptools. It warns you that it is depreacted but then it just hangs entirely16:56
clarkbI'll push up a gear change to fix that since I've sort it out anyway. Then we can land that change and the docker image update and then start planning to retire or hand it over16:56
opendevreviewJeremy Stanley proposed opendev/git-review master: Add release note for OSX Git version parsing fix  https://review.opendev.org/c/opendev/git-review/+/83855817:00
fungiclarkb: ^17:00
clarkbapproved thanks17:01
fungii guess after that merges, i'll tag 2.3.117:01
fungii need to disappear shortly, but can probably do that once i get where i'm headed in ~ an hour17:01
opendevreviewMerged opendev/git-review master: Publish documentation at every release  https://review.opendev.org/c/opendev/git-review/+/83834117:06
opendevreviewMerged opendev/git-review master: Add release note for OSX Git version parsing fix  https://review.opendev.org/c/opendev/git-review/+/83855817:23
opendevreviewClark Boylan proposed opendev/gear master: Fix python3.9 testing  https://review.opendev.org/c/opendev/gear/+/83856217:28
opendevreviewClark Boylan proposed opendev/gear master: Update the docker image to python3.9 and buster  https://review.opendev.org/c/opendev/gear/+/83840217:28
opendevreviewClark Boylan proposed opendev/storyboard master: Update container image to bullseye and python3.9  https://review.opendev.org/c/opendev/storyboard/+/83840317:31
fungiokay, i'm back on for a bit, tagging17:40
fungicommit e2382507ed080556c70b5166ed195f2edc05eae4 (HEAD -> master, tag: 2.3.1, origin/master, origin/HEAD, gerrit/master)17:43
fungioh, no that's not what we want. the other changes haven't merged yet17:43
funginevermind, it is what we want17:44
fungii forgot the last one was the release note change17:44
fungiclarkb: ^ does that look right to you?17:45
fungigit log --oneline --no-merges 2.3.0..2.3.117:46
fungi838558: Add release note for OSX Git version parsing fix17:46
fungi838341: Publish documentation at every release17:46
fungi838523: Fix get_git_version on OS X17:46
fungithat's what it covers17:46
opendevreviewClark Boylan proposed opendev/gear master: Fix python3.9 testing  https://review.opendev.org/c/opendev/gear/+/83856217:47
clarkbfungi: looking17:48
clarkbe2382507ed080556c70b5166ed195f2edc05eae4 as 2.3.1 lgtm17:48
fungiokay, pushing. fingers crossed we get a working doc update from it17:49
clarkbthat new docs job is running now17:50
fungilooks like it probably worked, need to wait for the timed vos release now17:55
fungiet voila! https://docs.opendev.org/opendev/git-review/latest/releasenotes.html17:55
fungii suppose that means we can go ahead and approve 838340 now too, which will refresh the bindep docs when it hits promote17:56
fungiand then we'll have updated release notes published for both bindep and git-review and can send out release announcements17:56
clarkbI approved it and rechecked it17:57
fungithanks! shouldn't need a recheck though17:57
clarkbI wasn't sure if zuul was smart enough to run it through check to get rid of the -117:57
fungiopendev tenant17:57
clarkbah right17:57
clarkb-1 doest' matter17:58
fungi99 problems, but a -1 ain't one17:58
fungisomething like that17:58
clarkbhttps://review.opendev.org/q/topic:opendev-image-cleanup+status:open everything in that list has +1s now. There are fixups for storyboard and gear testing as well as removal of unneeded buster container iamges and our python3.7 bullseye image18:07
clarkbOnce that has all landed I'll start to look at adding python3.10 images next18:07
opendevreviewMerged opendev/bindep master: Publish documentation at every release  https://review.opendev.org/c/opendev/bindep/+/83834018:08
clarkbfungi: if you have time to review https://review.opendev.org/c/opendev/system-config/+/838344 that would be good (refstack on the python 3.9 image). I can approve it after lunch today and keep an eye on it18:19
fungisure thing, i'll take a look once i'm done sending release announcements18:21
fungi#status log Released bindep 2.11.0 https://lists.opendev.org/pipermail/service-announce/2022-April/000035.html18:35
opendevstatusfungi: finished logging18:35
fungi#status log Released git-review 2.3.0 and 2.3.1 https://lists.opendev.org/pipermail/service-announce/2022-April/000036.html18:35
opendevstatusfungi: finished logging18:35
fungiclarkb: they all lgtm, approved the ones which could be18:46
opendevreviewMerged opendev/gear master: Fix python3.9 testing  https://review.opendev.org/c/opendev/gear/+/83856218:48
clarkbfungi: what about https://review.opendev.org/c/opendev/system-config/+/838344 ? (I didn't put it on the same topic as it isn't really in the path of removing the buster images)18:48
opendevreviewMerged opendev/gear master: Update the docker image to python3.9 and buster  https://review.opendev.org/c/opendev/gear/+/83840218:51
fungiahh, missed that one, thanks18:52
fungiyep, that one also lgtm18:53
opendevreviewMerged opendev/system-config master: Fixup some image dependencies in our zuul configs  https://review.opendev.org/c/opendev/system-config/+/83834518:53
opendevreviewMerged opendev/system-config master: Switch Refstack image over to python3.9  https://review.opendev.org/c/opendev/system-config/+/83834419:27
clarkbrefstack's conatiner updated 44 minutes ago and https://refstack.openstack.org/#/ is still serving content20:26
*** dviroel is now known as dviroel|out20:41
opendevreviewMerged opendev/storyboard master: Omnibus fixups for storyboard testing and CI  https://review.opendev.org/c/opendev/storyboard/+/83840921:34
opendevreviewMerged opendev/storyboard master: Update container image to bullseye and python3.9  https://review.opendev.org/c/opendev/storyboard/+/83840321:34
ianwif someone has a second to double-check my maths in https://review.opendev.org/q/topic:textarea-length that should stop 502 errors with long pastes on lodgeit21:57
ianwit's like working on an old car; somehow despite all the odds the thing keeps on trucking21:59
clarkbianw: one thing I'm confused about is if it is 64k bytes or characters22:00
clarkbsince they are different when it comes to unicode22:00
clarkbon the html side maxlength specifies characters not bytes22:00
clarkbthen on the python side we seem to be checking the length of a unicode string which is again characters not bytes22:01
clarkbbut on the mysql side I think it may be bytes?22:02
clarkbya that seems to be the case. That means this can still fail, but possibly less often?22:02
clarkbI think the first change is fine we can limit the size to a reaosnable upper bound. But the second one that checks the size should probably check the byte size not unicode length22:03
ianwi guess just .encode(utf-8)22:06
clarkbya I think so22:06
clarkband then len() of that value will by byte length22:06
opendevreviewIan Wienand proposed opendev/lodgeit master: new paste: check input length  https://review.opendev.org/c/opendev/lodgeit/+/83842122:08
ianwclarkb: another if/when you have time is https://review.opendev.org/c/openstack/diskimage-builder/+/837248 which is intended to be a no-op refactor of installing rpm-ish release packages.  over time that seems to have grown a bunch of add-on clauses that i think can be simplified22:09
ianwmostly looking for eyes on that to confirm it really is no-op :)22:09
clarkbcan do. I'm going to do a short garage bike ride then will take a look22:16
opendevreviewMerged openstack/project-config master: Remove tempest-lib from infra  https://review.opendev.org/c/openstack/project-config/+/83670322:31
clarkbianw: I agree it is a noop other than the centos stream ncurses stuff. I don't understand why they are brought in now after your change if they weren't prior to the explicit insallations23:04
clarkbleft a coupe of other minor things inline23:06
ianwi'm thinking because ncurses-libs wasn't pre-installed, so somehow ncurses-base gets chosen too because ... waves hands ... 23:07
ianwgreat point, i was thinking the centos path above matched it, but you're right, 7 falls through there23:08
opendevreviewIan Wienand proposed openstack/diskimage-builder master: yum-minimal: clean up release package installs  https://review.opendev.org/c/openstack/diskimage-builder/+/83724823:11
clarkbI'm going to approve https://review.opendev.org/c/opendev/system-config/+/838347 first thing tomorrow unless someone else wants to watch for image problems nowish23:23
clarkbI'm happy to do it and don't expect any problems now that gear and storyboard are happy23:23
clarkbI should approve the child too and remove the python3.7 bullseye image too. I'll also work on getting python3.10 bullseye images going23:23
opendevreviewIan Wienand proposed opendev/statusbot master: Prefix twitter logs with emoji  https://review.opendev.org/c/opendev/statusbot/+/83860823:29
ianwi'm happy to watch this23:30
opendevreviewIan Wienand proposed opendev/statusbot master: Prefix twitter logs with emoji  https://review.opendev.org/c/opendev/statusbot/+/83860823:37
ianwsometimes that alert icon looks like a yellow alert icon, and other times it looks like a black-and-white outline23:48
Clark[m]It rendered properly for me in the commit message and the python file diff on gerrit23:54
fungii could see it in gertty23:54
fungimy terminal font lacks the notice (pushpin) and ok (checkmark) glyphs though23:55
ianwahhh, til23:56
ianw'\u26A0\ufe0f' is different to '\u26A0\ufe0e'23:56
ianwthe last bit is a "selector" that says if it is a text or color version23:57
ianwif you leave that out, it may vary by platform and context which version is chosen23:57
fungineat. i knew about compositors for things like diacritics, but didn't realize there were also some for color selection23:58
opendevreviewMerged opendev/statusbot master: Prefix twitter logs with emoji  https://review.opendev.org/c/opendev/statusbot/+/83860823:58

Generated by irclog2html.py 2.17.3 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!