Tuesday, 2023-04-11

opendevreviewSteve Baker proposed openstack/diskimage-builder master: Allow custom console=tty0 argument  https://review.opendev.org/c/openstack/diskimage-builder/+/88000803:52
opendevreviewGhanshyam proposed openstack/project-config master: Retire openstack-health project: end project gating  https://review.opendev.org/c/openstack/project-config/+/88001304:52
opendevreviewSteve Baker proposed openstack/diskimage-builder master: Allow custom console=tty0 argument  https://review.opendev.org/c/openstack/diskimage-builder/+/88000805:15
opendevreviewTakashi Kajinami proposed openstack/project-config master: Retire puppet-rally - Step 5: Remove Project  https://review.opendev.org/c/openstack/project-config/+/88001906:14
*** amoralej|off is now known as amoralej06:29
*** jpena|off is now known as jpena07:05
*** elodilles_pto is now known as elodilles07:24
*** Trevor is now known as Guest1061108:42
*** amoralej is now known as amoralej|lunch11:57
*** dviroel_ is now known as dviroel12:14
*** amoralej|lunch is now known as amoralej13:01
corvusinfra-root: zuul-ci.org seems less than happy to me14:06
corvus#status log started apache on static01.opendev.org since it's serving zuul-ci.org14:07
opendevstatuscorvus: finished logging14:07
clarkbcorvus: zuul-ci.org shouldn't be served by static01 since friday. I only shutdown static01's apache yesterday after checking things looked good since friday14:41
clarkbcorvus: is it possible you've got stale dns somehow?14:41
clarkb(I also spot checked apache logs on static01 before shutting down apache there and it looked like crawlers were generating the traffic)14:42
clarkbin any case if you have more info I'll happily run that down as I'd like to get rid of static01 in the near future14:55
corvusclarkb: i can look into the dns possibility14:57
corvusclarkb: hrm, all of my dns now says it's static02.  so i'm sure it was a stale dns problem, but i no longer have the information needed to run it down (ie, i don't know if we had really long ttls, or my dns decided to serve a stale record)15:01
corvus#status log stopped apache on static01.opendev.org15:02
opendevstatuscorvus: finished logging15:02
clarkbttl should be 300 currently, however old ttl was 3600 and updated on friday15:02
clarkbI've got a couple changes proposed to bump the ttl back up again once we're happy with everything. I guess we can keep an eye on it for another day or two just to be extra careful given what you saw15:02
corvusit seems very likely that it was a local stale cache then (which is, weird).  sorry i lost the thread so i can't pull in it further.  :/15:03
clarkbit happens. I'll keep an ear out for any further trouble15:03
corvus(i didn't suspect that initially because other static.opendev.org sites worked for me)15:04
corvus(and despite having 28374927349387 terminals open, none of them are the one where i ran dig)15:05
clarkbI'm glad I'm not the only one who ends up with way too many terminals15:07
clarkbthe git 2.40 release notes taught me about git jump. I'm now hoping to find some time to figure that out as I think that may cut down on my terminal count15:09
clarkbthe 2.40 release just added emacs support too so that may be useful to you as well15:10
corvusthat looks pretty cool15:29
clarkbthe deploy job for the etherpad refactor was skipped due to letsencrypt job failing16:17
clarkbthe daily periodif job then succeeded. I'm going to double check on that before approving adding etherpad02 to the inventory16:17
clarkbok the periodic daily run of infra-prod-letsencrypt succeeded so not a consistent error whatever it ends up being16:18
clarkbI'll look at the job failure as soon as local updates are done and I can reboot and reload keys16:20
*** amoralej is now known as amoralej|off16:24
fricklerseems the new gerrit adds tags to links it includes in mails, like https://review.opendev.org/c/openstack/neutron-dynamic-routing/+/879337?usp=email16:30
clarkbfrickler: that might be editable in the email templates assuming they expose urls that don't have the tags?16:30
fricklerdo we evaluate that somewhere? otherwise maybe we can get rid of it to avoid giving people the impression we are tracking them somehow?16:30
clarkbwe don't use that anywhere, but gerrit internally might? for example gerrit does allow you to comment on reviews now via email and maybe that plays into it somehow?16:31
clarkb(note we don't have the mail server set up for that to work anyway aiui so not a problem for us)16:31
clarkbfrickler: maybe the first thing to do is ask upstream if that is important for something and see if maybe it can be changed upstream if not16:32
clarkbbasically flip it around and force the one or two installs that use it (chrome and android I would assume) to update their templates rather than the other way around16:32
fricklerclarkb: could you do that please? I've avoided the need to register in their bugtracker so far16:33
clarkbya though it might not be until later today. Doing a quick grep in the repo shows: https://gerrit.googlesource.com/gerrit/+/refs/heads/stable-3.7/polygerrit-ui/app/models/views/change.ts#5816:35
clarkband elsewhere usp is set to service-worker-notification16:36
clarkbI suspect that generally you can't filter all usp parameters. But it may be the case that the email one isn't helpful16:36
clarkbsometimes it is set to dashboard as well16:37
clarkbI don't think we can change this via the templates because the java around the templating is adding this to the url not the templates themselves https://gerrit.googlesource.com/gerrit/+/refs/heads/stable-3.7/java/com/google/gerrit/server/mail/send/ChangeEmail.java#25616:39
clarkbso I don16:41
clarkbI don't think this is about tracking for nefarious reasons but it does look like elsewhere they use this to inform web state since it is a webapp interacting with the api16:41
clarkbI don't see anywhere where email is used on the web ui side though. Some of he other stuff seems to be used though16:42
clarkbI think instead of a bug to start I'll write an email to the list asking for some more explanations of how it is used16:43
opendevreviewMerged openstack/diskimage-builder master: Rename openeuler mirror script  https://review.opendev.org/c/openstack/diskimage-builder/+/87880716:48
*** jpena is now known as jpena|off16:52
clarkbfrickler: https://groups.google.com/g/repo-discuss/c/kkbyIKkzlAU/m/4t789yerBAAJ16:59
clarkbmirror01.regionone.linaro.opendev.org acme.sh execution returned -1317:10
clarkbconsidering the next run was successful I think this is fine and I can probably approve the etherpad02 addition change17:11
clarkbhttps://review.opendev.org/c/opendev/system-config/+/879386 has been approved17:11
opendevreviewMerged opendev/system-config master: Add etherpad02 to inventory  https://review.opendev.org/c/opendev/system-config/+/87938617:35
ianwi was thinking on the wiki/issue tracking etc. i wonder if with gitea we could keep the load-balancing for git requests, but then have one big server to enable issue tracking/wiki19:53
ianwto try and avoid having it be a distributed system19:53
ianwi'm hand-waving the login requirements too, and things like gerrit integration closing issues19:54
clarkbI think if we do that we should deploy it without git repo content too19:55
clarkbI think it is theoretically possible, but would require some thought19:55
ianwit seems like, in theory, haproxy could be made to do it19:56
clarkboh i would decouple the url and fqdn too I think19:56
clarkbI mean yes what you are describing would be possible but we'd have to move all of the ssl termination into haproxy and I think it would make for a much more confusing setup operationally19:57
clarkbbut maybe having the opendev.org issues link for $project link to tracker.opendev.org as another similar looking gitea instance would be too confusing19:57
clarkber too confusing for endusers I mean19:58
clarkbdefinitely worthy of a spec if we go that route19:59
ianwi think a separate site would be painful if we did want to grow it into a cluster situation20:02
ianwit seems like proxying (HEAD|info/refs|objects/info/.*|git-upload-pack).* would be sufficient20:24
opendevreviewClark Boylan proposed opendev/system-config master: Remove static01 from inventory  https://review.opendev.org/c/opendev/system-config/+/88008620:30
opendevreviewClark Boylan proposed opendev/system-config master: Remove etherpad01 from inventory  https://review.opendev.org/c/opendev/system-config/+/88008720:30
ianwit also seems we could poc it on another port, although gitea-lb would need to get a cert20:30
clarkbI've WIP'd that second change but will leave the WIP off the first as I think we can remove that at any point once we are comfortable (tomorrow?)20:30
clarkbianw: on the etherpad you've noted centos 8 stream removal. cs8 isn't EOL until next year. Is there a reason to do that more quickly?20:33
ianwjust the practicality of keeping it going20:34
opendevreviewMerged openstack/diskimage-builder master: Check and grow the GPT structure to the end of the volume  https://review.opendev.org/c/openstack/diskimage-builder/+/82755820:35
ianwbut no, it doesn't need to be worked on right now20:35
clarkbok I tried to organize by topic on that etherpad with the idea being that interests could be used to steer people to applicable tasks and then hopefully use that to grow greater understanding of the whole (I forget how ianw put it in the meeting and I've just butchered it)20:39
ianwthanks, yep sorry it was a bit of a dump of my todo as is20:40
clarkbianw: heh re paste I think we were effectively upstream like 10 years ago20:54
ianwyeah, it hangs on, and i guess in some ways it's feature complete so it can just sit, but still20:56
ianwclarkb: can you make sense of this for NO_CHANGE matching?  https://gerrit-review.googlesource.com/c/gerrit/+/6463122:18
ianwif i'm understanding, you pushed a review for "refs/for/a-branch"; then "a-branch" merged back into master, then you rebase that change to master, but maintain your existing parent commit?22:27
opendevreviewIan Wienand proposed openstack/project-config master: wheel builds : move to individual releases  https://review.opendev.org/c/openstack/project-config/+/87972222:43
ianw^ was the result of : git commit --amend --no-edit --date "$(date)"22:44
ianwon the push i got22:45
ianw"remote: warning: no changes between prior commit ff367dc and new commit be56227"22:45
opendevreviewIan Wienand proposed openstack/project-config master: wheel builds : move to individual releases  https://review.opendev.org/c/openstack/project-config/+/87972222:47
opendevreviewIan Wienand proposed openstack/project-config master: wheel builds : move to individual releases  https://review.opendev.org/c/openstack/project-config/+/87972222:48
ianwopenstack/openstacksdk.config:[label "Allow-Post-Review"] is the only acl that didn't have a copycondition that got one23:19
ianwthe way it checks for migration is looking for a commit with the message "Migrate label configs to copy conditions"23:40

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