| *** yadnesh|away is now known as yadnesh | 04:17 | |
| *** jpena|off is now known as jpena | 08:40 | |
| ade_lee_ | fungi, clarkb https://review.opendev.org/c/zuul/zuul-jobs/+/874907/1 please | 10:56 |
|---|---|---|
| ade_lee_ | ianw, ^^ | 10:56 |
| ade_lee_ | looks like because the ua-registration-role is called by a playbook in project-config, zuul isn't allowing changes to be applied beforehand in this role | 10:58 |
| *** yadnesh is now known as yadnesh|away | 13:40 | |
| bauzas | fwiw, we currently face quite a number of jobs that fail due to ceph.com being flakey | 14:46 |
| bauzas | don't know yet whether it's massively impacting all changes or just a slice of them | 14:46 |
| bauzas | https://a2777df7406b272e83f9-872af65051a6414dfad04293989b5338.ssl.cf5.rackcdn.com/874515/3/check/nova-ceph-multistore/ba535b5/job-output.txt | 14:47 |
| bauzas | an opensearch query gives me a couple of occurrences across projects | 14:47 |
| fungi | if those are used frequently, we could look into adding a caching proxy for that site on our mirror servers | 15:08 |
| bauzas | looks like the transient issue is mostly gone | 15:09 |
| bauzas | but we'll see | 15:09 |
| dansmith | it's not gone for me | 15:41 |
| dansmith | fungi: nova, glance, devstack all runt he ceph job, FWIW | 15:41 |
| dansmith | so caching might be appropriate | 15:42 |
| dansmith | I can't say I remember it ever being a problem before, but right now download.ceph.com is a black hole for me, so 100% no merging for those projects (and maybe cinder, not sure) | 15:42 |
| dansmith | yeah, cinder also | 15:43 |
| fungi | mmm, slight flakiness can be improved by a caching proxy, but a complete outage would still be a problem regardless. we'd have to mirror copies of whatever's hosted there if that's what we need to guard against | 15:45 |
| dansmith | well, that's what I'm saying, I've never really seen it down, that I know of | 15:46 |
| dansmith | it's down now, but that's unusual, AFAIK | 15:46 |
| dansmith | anyway, you know better than I what is worthwhile, I'm just answering your "if used frequently" query | 15:47 |
| fungi | yep, that's good info | 15:50 |
| *** dulek_ is now known as dulek | 15:58 | |
| *** jpena is now known as jpena|off | 17:28 | |
| dtantsur | folks, could someone take a look at https://docs.openstack.org/api-ref/baremetal/ please? Am I the only who only sees very few headings? | 17:28 |
| dtantsur | same for compute, it seems. a regression in the api-ref job? a change in sphinx? | 17:28 |
| dtantsur | some CSS in our theme disables H1 headers using display:none. WUT. | 17:34 |
| dtantsur | https://github.com/openstack/openstackdocstheme/commit/f81f3344076a09482545534e014318d7e961f825 but it's quite old... | 17:36 |
| dtantsur | baremetal and compute are affected, network is not. puzzling. | 17:37 |
| dtantsur | yep, the former use <h1>, the latter - <h2> | 17:39 |
| JayF | https://raw.githubusercontent.com/openstack/neutron-lib/master/api-ref/source/v2/index.rst | 17:40 |
| JayF | I think neutron is adding an extra layer of headings? | 17:40 |
| dtantsur | aha! and this first layer is not displayed for them as well | 17:40 |
| dtantsur | e.g. there are "Networks" and no "Layer 2 Networking" | 17:41 |
| JayF | https://raw.githubusercontent.com/openstack/neutron-lib/master/api-ref/source/v2/address-groups.inc | 17:41 |
| JayF | yep | 17:41 |
| JayF | using that as an example, Address groups shows up | 17:41 |
| JayF | but show address groups / request / response do not | 17:41 |
| JayF | it looks like this is something we can fix the visuals of in Ironic proper; but I don't know which direction is the right one to go | 17:42 |
| dtantsur | okay, we can probably try to work around it in ironic.. but would be good if someone look at the openstack-wide impact | 17:42 |
| JayF | Best approach is probably to hit the list; would you like to or should I? | 17:43 |
| dtantsur | JayF: if you can write to the list, I can try a workaround | 17:43 |
| JayF | ack | 17:43 |
| dtantsur | JayF: one could argue, it's an actual mistake in the api-ref's, but we still need to make people aware | 17:44 |
| JayF | I don't care which direction is "right" (updating the theme or updating the content) | 17:44 |
| JayF | we just need to raise awareness of it since nova is impacted if it's in content | 17:44 |
| dtantsur | mm, the workaround is not exactly trivial. changing the top-level heading to something else fixes the subheading, but now sub-subheading are not in the navigation tree >_< | 17:48 |
| dtantsur | I need to think about it | 17:48 |
| fungi | JayF: dtantsur: that seems documentation related. odds are it's something with openstackdocstheme, in which case you want the #openstack-oslo channel | 18:00 |
| JayF | fungi: I hit the list about it; hopefully someone will reply there | 18:00 |
| fungi | (openstackdocstheme is an oslo deliverable) | 18:00 |
| JayF | makes sense | 18:01 |
| fungi | folks in here know very little about it for the most part | 18:01 |
| JayF | add me to that list ;) | 18:01 |
| JayF | I've hit the list, there's no further IRC chat to have about it for now I think, I'll see if someone knowledable responds there | 18:01 |
| fungi | there was something in a sphinx release a while back which changed how h1 headings are rendered/skipped, and there were changes to the theme to accommodate that, but if memory serves it also necessitated changes to the rst content. happened last year though i think? | 18:02 |
| JayF | That may be exactly what we (and nova) are hitting | 18:02 |
| fungi | no idea if there's anything more recent that's happened which might impact headings | 18:03 |
| JayF | it's exceedingly possible this has been broken a while and nobody noticed | 18:03 |
| fungi | let me see if i can find the openstackdocstheme change which was related to that sphinx behavior change | 18:03 |
| JayF | https://github.com/openstack/openstackdocstheme/commit/f81f3344076a09482545534e014318d7e961f825 added the specific css rule that seems to be hitting us now | 18:04 |
| JayF | but if they changed what things translate into h1/h2/etc in sphinx, that would've impacted it too, I'd assume | 18:04 |
| fungi | https://review.opendev.org/c/openstack/openstackdocstheme/+/798897 | 18:04 |
| fungi | that landed in openstackdocstheme 2.3.1 in late 2021 | 18:05 |
| fungi | that's the most recent thing i'm aware of around it | 18:05 |
| JayF | dtantsur: ^ look at fungi's https://review.opendev.org/c/openstack/openstackdocstheme/+/798897 link | 18:06 |
| JayF | I'm going to look for list posts around that time | 18:06 |
| fungi | good thinking, pretty sure there was an ml thread about it | 18:06 |
| fungi | maybe look at docs changes from that timeframe in projects not exhibiting the issue too | 18:07 |
| JayF | https://lists.openstack.org/pipermail/openstack-discuss/2021-November/025680.html | 18:07 |
| JayF | that looks sus | 18:07 |
| jjasghar | o/ | 22:58 |
Generated by irclog2html.py 2.17.3 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!