Monday, 2022-06-20

*** rlandy is now known as rlandy|holiday00:05
ianwinitial report on the login popup ->
ianwi searched and couldn't find a similar bug00:23
ianw is a potential fix that is running on the test server @
ianwit seems to work, but i have no idea if it's acceptable as a review, i very much hacked this from vague references in the javascript traceback of the 403 call and bits of git history01:48
ianwi.e. standard javascript development :)01:48
Clark[m]Looks like it failed some linting CI upstream01:56
ianwlooks to me like it's actually a job setup failure?
ianwCaused by: hudson.plugins.git.GitException: Command "/usr/bin/git fetch --tags --force --progress -- refs/changes/94/339494/1" returned status code 128:02:08
Clark[m]Maybe rerunning it is sufficient. I think there is a button you can click to do that02:24
*** pojadhav is now known as pojadhav|ruck04:30
opendevreviewyatin proposed zuul/zuul-jobs master: [DNM] Test proposal playbooks
opendevreviewyatin proposed zuul/zuul-jobs master: [DNM] Test proposal playbooks
opendevreviewyatin proposed zuul/zuul-jobs master: [DNM] Test proposal playbooks
*** pojadhav|ruck is now known as pojadhav|lunch07:02
opendevreviewyatin proposed zuul/zuul-jobs master: [DNM] Test proposal playbooks
*** yoctozepto[pto] is now known as yoctozepto07:14
*** yoctozepto is now known as Guest259307:15
opendevreviewyatin proposed zuul/zuul-jobs master: [DNM] Test proposal playbooks
*** jpena|off is now known as jpena07:36
*** yoctozepto_ is now known as yoctozepto07:50
*** pojadhav|lunch is now known as pojadhav|ruck07:52
opendevreviewyatin proposed zuul/zuul-jobs master: [DNM] Test proposal playbooks
opendevreviewyatin proposed zuul/zuul-jobs master: [DNM] Test proposal playbooks
*** rlandy_ is now known as rlandy|holiday09:54
*** pojadhav is now known as pojadhav|ruck09:55
*** dviroel_ is now known as dviroel12:05
fricklerwhile looking for log upload errors, I found these failures on ze01, maybe someone wants to have a closer look
fricklerhmm, did we change the merge conflict check in gerrit? it now sticks on a change that is definitively rebased
fricklercan't see it on the patch itself, but on the index
fungifor that first paste, i can't tell if the missing inventory key in the traceback is the reason the build was cancelled, or something that's happened because the job was cancelled (perhaps cancelled too early to have the node in its inventory? though that seems unlikely as well)12:35
*** ysandeep is now known as ysandeep|afk13:17
Clark[m]frickler: re the merge conflict thing that may be a cache problem since the info doesn't align across two different views.13:21
fungiit looked like it was claiming the change was parented to an abandoned change13:53
fungiahh, no i always read that table backwards. the abandon change has 836664 as its parent13:55
Clark[m]The mobile change listing doesn't show that info so not easy for me to confirm what frickler was seeing. But if I had to guess caches are involved13:56
fungithat seems likely13:56
fungii wonder if something didn't get updated during the reindex13:57
Clark[m]I wonder if we need to reindex more indexes. It only did changes automatically13:59
*** ysandeep|afk is now known as ysandeep14:00
fricklerthis seems somehow be related to WIP-ness, if I mark the patch as active, both WIP+Merge Conflict disappear in the index. once I mark it as WIP again, both are back14:08
fricklerand the issue persists even after pushing a new revision14:09
Clark[m]Interesting, that may be worth an upstream bug since it is tied to the WIP state14:10
Clark[m]I half wonder if all WIP changes are merge conflicted now for some reason14:11
Clark[m]Cannot merge because WIP ~= merge conflict maybe?14:11
fricklerthat would be an interesting logic. but indeed a random patch I mark WIP shows up with MC at once, too
ykarel_Clark[m], can you check
ykarel_and see which route can be taken to clear it15:14
ykarel_fungi, wrt, do you mean we need to handle setting that var on neutron side itself?15:18
ykarel_and other projects on their side?15:18
ykarel_something like
ykarel_or may be you meant something else15:19
Clark[m]ykarel re fips, yes you need to set it in your jobs15:20
Clark[m]And ya there isn't a good default because networking varies so much15:20
ykarel_Clark[m], okk i just see tempest unmerged patch also does the same
ykarel_ok will add on our side for neutron jobs15:22
*** ysandeep is now known as ysandeep|out15:29
ykarel_Clark[m], can you also check other one too
ykarel_just in case you missed15:35
*** dviroel is now known as dviroel|lunch15:36
*** marios is now known as marios|out16:03
Adri2000hello, how can I further debug a "Permission denied (publickey)" when using git-review? it used to work and I don't see what could have changed on my side recently...16:17
Adri2000trying to ssh manually to check if authentication work: ssh -p 29418 - I get the same error16:17
Clark[m]Adri2000: did you upgrade your operating system to include newer openssh? If so and you use rsa keys that is the likely issue16:18
Clark[m]Newer openssh dropped support for rsa + sha1 by default and will only do rsa + sha2 but gerrit's ssh server doesn't understand how to negotiate that in a way that makes openssh happy. This should be addressed when we upgrade to Gerrit 3.616:19
Clark[m]If you aren't using rsa keys then this may be a result of our recent Gerrit 3.5 upgrade, but this sounds like the rsa issue at first glance16:20
Adri2000ah, I upgraded to Ubuntu 22.04 recently, it's possible that I haven't tried pushing anything since then. and I use a 4096 rsa key indeed.16:21
Clark[m]There are two workarounds. First is to use an ed25519 or ecdsa key. The other is to set an ssh config to allow rsa + sha1. We aren't recommending this as it circumvents default security choices for your platform16:21
jrosser_could i get some help with this
jrosser_it doesnt appear to have run any jobs and i'm not seeing why that would be16:27
Adri2000Clark[m]: thank you, problem fixed using a new, ed25519, key :)16:34
*** dviroel|lunch is now known as dviroel16:43
*** jpena is now known as jpena|off16:45
clarkbjrosser_: typically the cause is file matches not matching. However, in this case it seems there is no stable/yoga branch defined for openstack/openstack-ansible and it appears to supply the job definitions for the templates on that project. I think that may be the reason16:45
jrosser_aaah ok16:45
jrosser_interesting - as we generally have to branch all these ansible roles first in order to then reference them when creating the yoga branch in openstack-ansible16:46
jrosser_noonedeadpunk: ^16:46
jrosser_clarkb: thanks for the pointer on missing branches... recheck has kicked the jobs off, it was probably all very racey with creating the branches and then finding the jobs across so many repos17:21
clarkbyou're welcome17:21
opendevreviewyatin proposed openstack/project-config master: Only install python3.8-venv in translation update jobs
ykarel_Thanks Clark[m] done ^17:24
ykarel_fungi, frickler  you too please check when get chance17:26
ykarel_thx in advance17:27
* ykarel_ out17:27
*** dviroel is now known as dviroel|afk20:05
*** rlandy|holiday is now known as rlandy|out20:16
*** dviroel|afk is now known as dviroel|out21:05
clarkbianw: upstream gerrit gave you the votes you need to merge your login required fix. They let the change authors self submit once you have the necessary approvals. Its great to see them review that so quickly22:36
clarkbthen we can trigger a rebuild of our images to pull in the fix and deploy it22:40
clarkblast call for meeting agenda items. I'm about to send it out22:46
opendevreviewMerged openstack/project-config master: Only install python3.8-venv in translation update jobs
clarkband sent22:58
ianwclarkb: yeah, will do today, i guess needs to forward port too23:11
clarkbianw: ya once you merge the change to 3.5 you will want to merge the resulting 3.5 into 3.6 and push that then merge 3.6 into master and push that. If you want to be lazy usptream does do those periodically to pick up changes that people didn't do them for. Half the time I try to be nice and push those forward ports they end up doing it themselves anyway (to pick up even more changes23:13
clarkbthat have landed since I pushed my changes)23:13
ianwlooks like it has a conflict on 3.6 -- back to the javascript :)23:19
ianwseems it has had a bit of a rewrite, so there's a new fn setReviewedStatus() that looks like it probes for logged in status.  also, it doesn't do it on (3.6.1)23:23
ianwwe can double check it as we get 3.6 in our ci23:24
fungiso maybe this is something they (accidentally?) fixed in 3.623:24
ianwfungi: yep, was part of other work, but i think did also fix this issue23:26
clarkbianw: I closed your issue23:33
clarkb(at some point my gerrit issue tracker perms got bumped up)23:33
ianwprobably when you started showing an aptitude for enterprise java development :)  thanks23:35
fungithat sounds terrifying23:42
opendevreviewIan Wienand proposed opendev/system-config master: gerrit: trigger rebuild for anonymous browsing fix

Generated by 2.17.3 by Marius Gedminas - find it at!