Wednesday, 2025-09-03

corvus_https://bsky.app/profile/did:plc:2uvuwgm6digf5tjoi3iwkxcz/post/3lxuslbzjuc2t is more info on the matrix.org issues00:46
opendevreviewOpenStack Proposal Bot proposed openstack/project-config master: Normalize projects.yaml  https://review.opendev.org/c/openstack/project-config/+/95799502:51
mnasiadkacorvus_: 55TB database is interesting06:44
*** dhill is now known as Guest2575210:52
*** dhill is now known as Guest2575512:02
opendevreviewTadas Sutkaitis proposed openstack/diskimage-builder master: feat: introduce new cloud-init service names  https://review.opendev.org/c/openstack/diskimage-builder/+/95931612:32
opendevreviewTadas Sutkaitis proposed openstack/diskimage-builder master: feat: introduce new cloud-init service names  https://review.opendev.org/c/openstack/diskimage-builder/+/95931612:38
opendevreviewTadas Sutkaitis proposed openstack/diskimage-builder master: feat: introduce new cloud-init service names  https://review.opendev.org/c/openstack/diskimage-builder/+/95931612:39
corvus_i'm restarting the schedulers since the memory fix merged, and they are starting to swap.13:19
corvus_okay, that's done13:31
fricklercorvus_: did that only affect the schedulers or were they only the most affected services?13:48
corvus_should be just schedulers (and maybe web).  i restarted both.13:49
fricklerok, thx13:52
clarkbthe last post on bluesky says they think they will be up 3-4 hours from the post and the post is about 3 hours old14:44
clarkbhere's hoping the estimate is accurate14:44
clarkbjrosser: I think the debian trixie image should be actually trixie now if you want to confirm14:49
jrossersure, i can recheck some things and see how it goes14:49
opendevreviewTadas Sutkaitis proposed openstack/diskimage-builder master: feat: introduce new cloud-init service names  https://review.opendev.org/c/openstack/diskimage-builder/+/95931615:18
opendevreviewTadas Sutkaitis proposed openstack/diskimage-builder master: feat: introduce new cloud-init service names. Script logic inspired by: https://github.com/openstack/diskimage-builder/commit/93442d09cff0eadfc18b17793c24789dac094600  https://review.opendev.org/c/openstack/diskimage-builder/+/95931615:20
opendevreviewMohammed Naser proposed openstack/diskimage-builder master: feat: introduce new cloud-init service names.  https://review.opendev.org/c/openstack/diskimage-builder/+/95931615:23
clarkbhttps://review.opendev.org/c/openstack/project-config/+/957995's diff makes firefox sad15:54
clarkblooks like it is linewrapping text blocks differently to before (shorter line widths and now with extra spaces)15:55
clarkbfrickler: you -1'd it in the past not sure if that was for the same linewrapping thing. I went ahead and +2'd it because I think it is valid and I'm not sure I want to fight ruamel or whatever is normalizing things. But feel free to -1 again if you prefer. I can drop my +215:56
clarkbrackspace sent an email a few minutes ago indicating the maintenance affecting the sjc3 mirror volume is completed and we can turn things back on. I'm actually going to pop out for a bike ride now before it gets hot today and then I'll reenable that mirror and the region in zuul when I get back16:48
fricklerclarkb: commented with -116:53
clarkbfrickler: ack I pulled my +216:58
corvusclarkb: hello from matrix17:01
corvus_bridge lag is about 1 minute right now17:01
clarkbhello!17:02
clarkb"The database has been restore, and verification checks complete. matrix.org's Synapse has been started up again and we are monitoring"17:04
clarkbok popping out now. Back in a bit17:04
corvus_ooh bridge lag is down to a couple of seconds now17:04
opendevreviewMerged openstack/diskimage-builder master: feat: introduce new cloud-init service names.  https://review.opendev.org/c/openstack/diskimage-builder/+/95931617:42
*** Guest22851 is now known as diablo_rojo_phone18:12
opendevreviewMerged openstack/diskimage-builder master: Drop remaining reference to TripleO  https://review.opendev.org/c/openstack/diskimage-builder/+/95888418:28
nhicher[m]clarkb: hello, quick question about multi-node-bridge role on zuul-jobs. Is it ok if I drop everything related to ovs and have geneve/vxlan or do you prefer I'll keep ovs bits?18:45
Clark[m]nhicher: I think the idea was to drop ovs due to Ansible removing the module for it. Then rely on Linux bridges which are more common without supporting tooling18:53
Clark[m]So it is less about geneve and more about replacing ovs with something more common18:54
opendevreviewMerged openstack/diskimage-builder master: Remove mirror from experimental debian jobs  https://review.opendev.org/c/openstack/diskimage-builder/+/95325618:54
opendevreviewMerged openstack/diskimage-builder master: Fix debian-systemd element for non-root user  https://review.opendev.org/c/openstack/diskimage-builder/+/95287518:54
nhicher[m]<Clark[m]> "So it is less about geneve and..." <- yes, I'm working with iproutes2 and brctl, my question is should I keep ovs part for now, then remove it later. For me it's easier to clean everything related to ovs. I will keep the same role name 'multi-node-bridge' to avoid to have something to change on users side19:03
opendevreviewMerged openstack/diskimage-builder master: Add support for DNF5-based systems  https://review.opendev.org/c/openstack/diskimage-builder/+/93433219:03
clarkbnhicher[m]: I think as long as we can test that the new thing works broadly there isn't any reason to keep ovs19:15
clarkband we should be able to do that with zuul-jobs test framework without too much trouble19:15
opendevreviewClark Boylan proposed opendev/zuul-providers master: Revert "Disable raxflex sjc3 for cloud maintenance window"  https://review.opendev.org/c/opendev/zuul-providers/+/95937919:20
opendevreviewMerged opendev/zuul-providers master: Revert "Disable raxflex sjc3 for cloud maintenance window"  https://review.opendev.org/c/opendev/zuul-providers/+/95937919:21
opendevreviewDr. Jens Harbott proposed openstack/project-config master: Switch devstack propose-updates job to noble  https://review.opendev.org/c/openstack/project-config/+/95938219:28
clarkbsjc3 is back to doing work again according to grafana19:43
clarkbcorvus: it looks like the zuul launcher nodes listing does not provide the cloud side uuid or name which are useful if you want to lookup the node in the cluod. I think the zuul-launcher id does show up as a prefix in the node name though?19:53
corvusclarkb: yep.  we'd need a little bit of work to plumb that info through to the ui.  but in the interim, looking for a node with the first few chars of the id is reliable19:55
clarkbcool thanks for confirming19:56
corvuswill be "np...."19:58
corvus(because windows can't handle numeric hostnames19:58
clarkbsounds like the matrix.org issues started with some planned storage updates/testing? oof20:08
clarkbthey made a post promising a full postmortem eventually. That will be a fund read I'm sure20:08
opendevreviewMerged openstack/diskimage-builder master: Add a sha256sum check for CentOS Cloud Images  https://review.opendev.org/c/openstack/diskimage-builder/+/95798320:17
opendevreviewNicolas Hicher proposed zuul/zuul-jobs master: Refactor: multi-node-bridge to use linux bridge  https://review.opendev.org/c/zuul/zuul-jobs/+/95939321:36
clarkbfrickler: I think https://sourceforge.net/p/ruamel-yaml/code/ci/375c4db4c7c5fad61e149fc0690e50229df23ab3/tree/emitter.py?diff=7218ee8b837cb2178b5910b5c0d50ab932a5eaed is the source of the ruamel trailing space problem. I almost wonder if we line wrapped ourselves if it would leave things alone (I think it would). Alternatively I think we can change the width to some large value and21:59
clarkbavoid this 99.99% of the time21:59
clarkbI think I actually like that myself. I'll push a patch21:59
opendevreviewClark Boylan proposed openstack/project-config master: Disable ruamel automagic line wrapping  https://review.opendev.org/c/openstack/project-config/+/95939622:08
clarkbfrickler: ^ something like that maybe22:08
opendevreviewNicolas Hicher proposed zuul/zuul-jobs master: Refactor: multi-node-bridge to use linux bridge  https://review.opendev.org/c/zuul/zuul-jobs/+/95939323:23

Generated by irclog2html.py 4.0.0 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!