| Clark[m] | Ack sorry I got roped into dinner prep myself | 00:27 |
|---|---|---|
| Clark[m] | tonyb: it's based on the build time of the image not when we update the image on the host so it is imperfect | 00:27 |
| tonyb | ahhh okay that makes sense. | 00:54 |
| *** mmalchuk_ is now known as mmalchuk | 08:40 | |
| *** lucasagomes_ is now known as lucasagomes | 12:08 | |
| AlbinVass[m] | if you don't me asking. What are the hardware specs for the machine opendevs gerrit instance is running on? | 12:30 |
| fungi | AlbinVass[m]: virtual machine running in vexxhost's ca-ymq-1/yul (montreal) region: 32 CPUs (AMD EPYC-Rome 3GHz), 128GB RAM | 12:35 |
| AlbinVass[m] | fungi: thanks! | 12:36 |
| fungi | overpowered at this point probably, but there was a time when we were coping with memory leaks and inconsiderate crawlers overloading it with resource-intensive queries | 12:37 |
| fungi | so the extra capacity was helpful | 12:37 |
| AlbinVass[m] | fungi: iirc you don't use gerrit mirrors right? | 13:29 |
| tonyb | AlbinVass[m]: correct single Gerrit no HA or mirroring | 13:59 |
| Clark[m] | But we do replicate to a fleet of gitea servers where we try to push the bulk of git traffic. The Gerrit is primarily dealing with people pushing code and CI systems fetching. | 14:42 |
| elodilles | hi infra, note that i'm about to do a branch clean up (deleting stable/zed branches that got tagged with *zed-eom*), as far as i see things are in a good shape for that | 15:24 |
| Clark[m] | elodilles: I haven't quite woken up yet today but ya I don't see or know of any reason not to do that | 15:25 |
| tonyb | elodilles: go for it. I'm around if you need anything | 15:25 |
| elodilles | Clark[m] tonyb : ACK, thanks o/ starting the script then | 15:30 |
| elodilles | and these branches were deleted this time: https://paste.opendev.org/show/bOF7ugF0UNdshesqZFSF/ | 15:50 |
| elodilles | tonyb: fyi, requirements, devstack*, grande were left out intentionally from this run, too | 15:50 |
| tonyb | elodilles: Okay cool. | 16:28 |
| tonyb | but requirements:stable/zed has been tagged eom so if anything lands in that review queue I should just abandon it, correct? | 16:29 |
| opendevreview | Clark Boylan proposed opendev/system-config master: DNM Forced fail on Gerrit to test the 3.9 upgrade https://review.opendev.org/c/opendev/system-config/+/893571 | 16:39 |
| clarkb | I rebased that on the docs build chnage because I'd really like to have 3.8 and 3.9 built that way if possible. I'll be able to check the docs more directly this way to see if building without webfonts presents issues (potential format problems maybe?) | 16:40 |
| clarkb | then can use the same node to test upgrade stuff since the changes are largely superficial and shouldn't impact the upgrade process | 16:41 |
| tonyb | clarkb: cool. Are you still thinking you'll propose the patch upstream? | 16:42 |
| clarkb | tonyb: yes. My hope is that we don't have to carry a downstream patch at all | 16:43 |
| clarkb | this is me abusing the CI system to avoid setting up bazelisk locally. It helps that our CI jobs populte the server with useful thinsg too (though in this case the update being docs related means that isn't really necessary) | 16:44 |
| tonyb | elodilles: FWIW, I ran a sync after your deletes and it all looks good locally. | 16:46 |
| tonyb | elodilles: I did some spot checking in zuul and the branches seem to have gone there too so no missed events | 16:46 |
| tonyb | clarkb: I don't see it as abuse, I see it as efficient | 16:47 |
| opendevreview | Merged opendev/system-config master: Remove some old CentOS compatibility and address FIXMEs https://review.opendev.org/c/opendev/system-config/+/918016 | 16:48 |
| tonyb | clarkb: diablo_rojo and I did some testing on the rebased meetpad changes and apart from certificate issues it looks good. | 16:49 |
| tonyb | clarkb: there's a held node if you're interested in checking | 16:49 |
| clarkb | sure. Do I have to edit /etc/hosts? | 16:52 |
| tonyb | Yup let me grab the entries | 16:53 |
| tonyb | # Temp hosts for OpenDev testing | 16:53 |
| tonyb | 104.130.4.30 meetpad99.opendev.org | 16:53 |
| clarkb | I'm in https://meetpad99.opendev.org/isitbroken | 16:55 |
| clarkb | it seems to be working | 16:55 |
| clarkb | gerrit 3.8 webfont free: https://158.69.71.178/Documentation/index.html vs our current docs: https://review.opendev.org/Documentation/index.html there is a difference for me but its small and I think workable | 17:50 |
| clarkb | and firefox web tools confirm we're no longer loading webfonts so it actually is doing what we want there | 17:52 |
| clarkb | so ya I guess the next step is pushing this code upsteram | 17:52 |
| tonyb | They looks the same to be in both chrome and firefox | 17:53 |
| clarkb | tonyb: you may have more fonts than I do or different font settings so you don't have to fallback to defaults? In my case the webfont free text is slightly larger | 17:54 |
| clarkb | and I think the bolds are more bold in the developer docs page for example | 17:54 |
| clarkb | but all of that is super minor and I think worth the benefit of not needing webfronts | 17:54 |
| tonyb | and yup I can confirm that I'm not downloading the fonts from the held node | 17:54 |
| clarkb | I'm figuring out how to push to upstream gerrit again so I can push this patch up to them and see what they say about the approach | 17:55 |
| tonyb | Yeah okay I can see that the bold is more bold on https://158.69.71.178/Documentation/linux-quickstart.html vs https://review.opendev.org/Documentation/linux-quickstart.html | 17:55 |
| elodilles | tonyb: yepp, just abandon it. i really hope that we don't need to do any further final stable zed releases (mostly inactive projects' stable/zed branch remained open, but i hope we can close them ~next week) | 18:00 |
| elodilles | tonyb: + thanks for checking the deletions | 18:01 |
| tonyb | elodilles: all good | 18:01 |
| tonyb | clarkb: keycloak01 is made obselete by keycloak03 correct? We just haven't removed it because paranoid? | 18:03 |
| clarkb | tonyb: ya I think cleanup of that server is on fungi's todo list | 18:07 |
| tonyb | Okay. | 18:08 |
| tonyb | It showed up in my distro audit | 18:08 |
| fungi | not around at the moment (just checked into a hotel for the night), but yes, i wanted to check the previous server's openinfraid integration before i delete it | 18:09 |
| tonyb | fungi: All good. No rush / pressure | 18:09 |
| clarkb | remote: https://gerrit-review.googlesource.com/c/gerrit/+/424404 Add release war target without webfonts | 18:10 |
| tonyb | Nice. | 18:12 |
| clarkb | I claerly haev a bias in that I push changes to our gerrit almost every day and help maintain git-review, but I have to say it is much much easier to push to our gerrit than upstream's gerrit | 18:14 |
| clarkb | for upstream you have to click scary buttons saying you're giving googlesource.com write access to all your git repos then they spit out a script to write cookies into a git cookie file (whcih I didn't run but instead manually wrote out because paranoia) then you also have to manually get the commit hook script installed somehow | 18:14 |
| tonyb | Wow, that's kinda gross | 18:15 |
| clarkb | and it isn't entirely clear if you push to the gerrit server url or the git server url (I pushed to the git server url and that worked, not sure if the alternative works). Anyway long story short there is a lot more to figure out and "solve" to push upstream | 18:15 |
| clarkb | another interesting thing with that change is upstream gerrit has a submit requirement requiring no unresolved comments to merge | 18:25 |
| clarkb | I don't think we should enforce that in our gerrit as that is quite rigid and trivial to overcome (anyone can resolve a thread iirc) | 18:25 |
| tonyb | Yeah I don't think that's fair. either | 18:28 |
| tonyb | How much of the awkward could be helped by adding some support to git review? | 18:29 |
| tonyb | Here's the very very first pass at the list of Focal -> {Jammy, Noble} updates. | 18:35 |
| clarkb | its possible that git review would already alleviate some of the weirdness, but you'd have to know the proper way to configure it for their server and https specific cookie stuff | 18:45 |
| clarkb | especially now that git review vendors the commit message hook the httpness of their server isn't as big a problem for that | 18:45 |
| tonyb | ... and this time I'll actually post the etherpad: https://etherpad.opendev.org/p/opendev-focal-server-upgrades | 18:49 |
| tonyb | clarkb: Yeah I guess something to add to the "if there was time" list :) | 18:51 |
| clarkb | tonyb: for the classification do you want me to move things or should I just label them? | 19:45 |
| clarkb | I went ahead and put some labels on things as that is least destructive (you can alwyas just remove the labels) | 19:48 |
| fungi | i saw the zuul-discuss post about options to respect the "no unresolved comments" rule in gerrit, i agree that seems... rigid | 20:08 |
| JayF | clarkb: I'll keep that in mind. I wonder if some of the negative feelings I've experienced around are you some Geritt might be bad experiences from other places | 21:08 |
| clarkb | JayF: ya I mean I semi regularly push to github and our gerrit and I think even gitlab once or twice. I don't think ours is mechanically much different. You put a pubkey in your account and run a tool to push. Upstream's gerrit is just very googley and has extra bits that are painful | 21:18 |
| JayF | Ah so when you say upstream, you literally mean the server to contribute to the gerrit project not the default out of box configuration | 22:03 |
| clarkb | yes | 22:03 |
| JayF | Okay that was the entire root of my misunderstanding. | 22:04 |
| opendevreview | Merged openstack/project-config master: Add new repository for NebulOuS testing data https://review.opendev.org/c/openstack/project-config/+/916876 | 22:08 |
| opendevreview | Goutham Pacha Ravi proposed openstack/project-config master: Adjust max Code-Review possible on governance repo https://review.opendev.org/c/openstack/project-config/+/918146 | 22:15 |
| opendevreview | Goutham Pacha Ravi proposed openstack/project-config master: Remove driverfixes branch acls from manila https://review.opendev.org/c/openstack/project-config/+/918147 | 22:21 |
Generated by irclog2html.py 2.17.3 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!