| @clarkb:matrix.org | infra-root if anyone has time to weigh in on https://review.opendev.org/c/opendev/system-config/+/973540 to test lodgeit backups over tcp rather than unix socket that would be appreciated. Also there is the etherpad 2.6.0 upgrade here: https://review.opendev.org/c/opendev/system-config/+/972001. | 16:00 |
|---|---|---|
| @clarkb:matrix.org | fungi: you had previously reviewed the gerrit 3.12 stack https://review.opendev.org/c/opendev/system-config/+/973970 is a newer change based on some feedback to reenable the 3.11 jobs independently of the nodejs 24 change you have previously reviewed. I think that one can go in today | 16:00 |
| -@gerrit:opendev.org- Takashi Kajinami proposed: | 16:26 | |
| - [openstack/diskimage-builder] 974342: Declare Python 3.11/12 support https://review.opendev.org/c/openstack/diskimage-builder/+/974342 | ||
| - [openstack/diskimage-builder] 974343: Bump minimum python version to 3.8 https://review.opendev.org/c/openstack/diskimage-builder/+/974343 | ||
| -@gerrit:opendev.org- Takashi Kajinami proposed: [openstack/diskimage-builder] 974344: Drop unnecessary __future__ import https://review.opendev.org/c/openstack/diskimage-builder/+/974344 | 16:28 | |
| -@gerrit:opendev.org- László Jánosi proposed: [opendev/glean] 974346: Fix vlan and bond config generation https://review.opendev.org/c/opendev/glean/+/974346 | 17:10 | |
| -@gerrit:opendev.org- Stephen Finucane proposed: [openstack/project-config] 974347: Clean up ACLs for SDK projects https://review.opendev.org/c/openstack/project-config/+/974347 | 17:23 | |
| @clarkb:matrix.org | fungi: then if you or someone else is willing to help I think we can alnd the nodejs 24 change first thing tomorrow and restart Gerrit on that new build. I have some errands to run in the afternoon (school pickup so still mostly around) which means it would be good to get that done early in the day | 18:24 |
| @fungicide:matrix.org | yeah, i'm around all day tomorrow. making pizza dough at some point but that's about as complicated as my day's getting so should be fine | 18:25 |
| @clarkb:matrix.org | ooh pizza | 18:26 |
| @fungicide:matrix.org | request for more reviews on https://review.opendev.org/c/opendev/system-config/+/974243 as once again lists.o.o is taking forever to load the web interface for me | 18:31 |
| @clarkb:matrix.org | fungi: I think we can probably go ahead and approve that one with my review on it | 18:32 |
| @clarkb:matrix.org | it is an easy undo if it comes to that so the risk is low | 18:32 |
| @clarkb:matrix.org | I approved it | 18:33 |
| @fungicide:matrix.org | thanks. getting "The server is temporarily unable to service your request due to maintenance downtime or capacity problems. Please try again later." from it right now sadly | 18:34 |
| @clarkb:matrix.org | fungi: if you want you can manually apply the change and reload apache2 to look at it quicker | 18:35 |
| @clarkb:matrix.org | I do wonder if we simply need to add more slots to the server (though that is trickier to do with this service than it was with static) | 18:35 |
| @fungicide:matrix.org | i'm not in a huge hurry, was just trying to pull some posts up from a list archive to copy some information that i was too busy to dig out of my sent mail | 18:36 |
| @clarkb:matrix.org | also I'm going to try and get out for a bike ride this afternoon (shouldn't really impact anything just a heads up) | 18:37 |
| @fungicide:matrix.org | great idea | 18:41 |
| @mnaser:matrix.org | this seems very strange, where is this coming from? | 19:08 |
| @fungicide:matrix.org | that url just gives me `{"errcode":"M_NOT_FOUND","error":"Not found"}` | 19:09 |
| @fungicide:matrix.org | maybe that's an error coming back from matrix.org | 19:09 |
| @mnaser:matrix.org | ah, that was an image i uploaded | 19:09 |
| @fungicide:matrix.org | might be my console client for matrix isn't returning the correct url | 19:09 |
| @mnaser:matrix.org | https://i.imgur.com/XXFztG6.png plain o'l image upload | 19:09 |
| @mnaser:matrix.org | tl;dr visiting codesearch.opendev.org in chrome tells me that the site wants to "Look for and connect to any device on your local network" | 19:10 |
| @fungicide:matrix.org | um | 19:10 |
| @fungicide:matrix.org | yeah that's definitely odd and not something i'd expect | 19:10 |
| @mnaser:matrix.org | weird.. opened console, refreshed a few times, no longer doing it | 19:11 |
| @fungicide:matrix.org | client-side malware? three-letter agency mitm your internet? | 19:11 |
| @mnaser:matrix.org | i mean, i run pi.hole at home, i wonder if it was resolving something to a local ip and it didnt like it | 19:12 |
| @fungicide:matrix.org | looking in the page source there's a `<link rel="search" href="//localhost:6080/open_search.xml" type="application/opensearchdescription+xml" title="Hound" />` which seems kinda broken/useless but i doubt that's the cause | 19:12 |
| @mnaser:matrix.org | oh i think that might very well be it potentially if its trying to make a request to the local network | 19:13 |
| @fungicide:matrix.org | localhost:6080 is the socket apache is proxying https requests to, so i think hound is a little confused in including that search link | 19:13 |
| @mnaser:matrix.org | yep, i think so too, its mean to help browsers "autoregister" search engines it seems (application/opensearchdescription+xml) | 19:14 |
| @mnaser:matrix.org | so i think chrome might have been trying to scrape that and before it can do that it needs to ask for permissions | 19:14 |
| @fungicide:matrix.org | i wouldn't expect a browser to follow that link, but maybe some do try to prefecth anything in a link tag | 19:14 |
| @fungicide:matrix.org | er, prefetch | 19:14 |
| @mnaser:matrix.org | to be fair, this is also broken - https://codesearch.opendev.org/open_search.xml | 19:15 |
| @mnaser:matrix.org | think its localhost in ther etoo | 19:15 |
| @fungicide:matrix.org | yeah, i honestly have no idea what that's for even, and it'll take someone digging to figure it out and how to correct or disable it | 19:15 |
| -@gerrit:opendev.org- Zuul merged on behalf of Clark Boylan: [opendev/system-config] 973970: Reenable Gerrit 3.11 image builds https://review.opendev.org/c/opendev/system-config/+/973970 | 19:19 | |
| -@gerrit:opendev.org- Zuul merged on behalf of Jeremy Stanley https://matrix.to/#/@fungicide:matrix.org: [opendev/system-config] 974243: Don't rewrite the mod_status path in Apache https://review.opendev.org/c/opendev/system-config/+/974243 | 19:19 | |
| @clarkb:matrix.org | it is interesting that firefox doesn't do that | 19:19 |
| @clarkb:matrix.org | firefox never makes a request to that url so either firefox internally decides that it shouldn't request localhost resources or it never follows the url down that path | 19:21 |
| @clarkb:matrix.org | as expected gerrit 3.11 did not get a promotion job due to 973970 landing (whcih is what we wanted) | 19:22 |
| @clarkb:matrix.org | fungi: mnaser https://github.com/hound-search/hound/blob/main/ui/assets/index.tpl.html and https://github.com/hound-search/hound/blob/main/ui/assets/open_search.tpl.xml indicate that the .Host variable supplies that value wherever it comes from | 19:26 |
| @clarkb:matrix.org | and it has been there for forever | 19:27 |
| @clarkb:matrix.org | https://github.com/hound-search/hound/blob/main/ui/ui.go#L162 I think that value comes from here. Which looks like it is using the http requset info. Maybe if we set x forwarded for it would work properly? | 19:30 |
| @clarkb:matrix.org | and I'm guessing the new behavior is due to some update to chrome making it smarter about these urls? | 19:30 |
| @clarkb:matrix.org | fungi: I think lists' apache should be updated now assuming we issued a reload for that (I think we do?) | 19:33 |
| @fungicide:matrix.org | yep, it's working now | 19:36 |
| @mnaser:matrix.org | I just did update Chrome though so maybe its a new thing in a new version or something, but I can look at the code too in a minute | 19:37 |
| @clarkb:matrix.org | We want X-Forwarded-Host not -For and then assuming the golang http request object does the right thing the host value will be updated? I'm actually thinking that may not work because the template call explicitly uses Host | 19:44 |
| @clarkb:matrix.org | https://cs.opensource.google/go/go/+/refs/tags/go1.25.6:src/net/http/request.go;l=226 yup I don't think setting the x forwarded host will help. This is a bug in hound imo | 19:47 |
| @clarkb:matrix.org | I don't see an upstream bug yet, but also upstream is in very maintenance mode | 20:01 |
| @clarkb:matrix.org | May be worth filing just to make sure we aren't doing something silly | 20:02 |
| @mnaser:matrix.org | i think you can just send the Host header directly to the backend, no ? | 20:03 |
| @mnaser:matrix.org | like "proxy_set_header Host codesearch.opendev.org" | 20:04 |
| @clarkb:matrix.org | Oh maybe. I assume mod proxy sets it but that may be overrideable | 20:05 |
| @clarkb:matrix.org | X-Forwarded-Host is the correct way though | 20:05 |
| @mnaser:matrix.org | I agree with you, that is the correct way, but I _suspect_ that Hound is not aware of running behind an HTTP server | 20:05 |
| @mnaser:matrix.org | Clark: https://httpd.apache.org/docs/current/mod/mod_proxy.html#proxypreservehost | 20:06 |
| @clarkb:matrix.org | Right hound is buggy | 20:06 |
| @clarkb:matrix.org | I can write a change to preserve host | 20:07 |
| @mnaser:matrix.org | ill push that up real quick :) | 20:08 |
| @clarkb:matrix.org | Can you add a testinfra test to check the resulting html? | 20:09 |
| @clarkb:matrix.org | Or I can as a follow up | 20:09 |
| @clarkb:matrix.org | https://opendev.org/opendev/system-config/src/branch/master/testinfra/test_codesearch.py#L28 can update this test | 20:10 |
| @clarkb:matrix.org | Assert localhost not in cmd.stdout or similar | 20:10 |
| @mnaser:matrix.org | thanks that link is helpful | 20:11 |
| @mnaser:matrix.org | i will add it there too as a check | 20:11 |
| @clarkb:matrix.org | Thanks | 20:11 |
| @fungicide:matrix.org | we can certainly benefit from more validation checks on a lot of the less-exciting services we run | 20:14 |
| @mnaser:matrix.org | codesearch is a very exciting useful one for me :) | 20:14 |
| @fungicide:matrix.org | i agree it's useful, but it's one of those things people tend to forget about again as soon as they're done making a query | 20:15 |
| @fungicide:matrix.org | more part of the boring plumbing of running a code hosting community | 20:16 |
| -@gerrit:opendev.org- Mohammed Naser proposed: [opendev/system-config] 974365: Add ProxyPreserveHost to codesearch vhost https://review.opendev.org/c/opendev/system-config/+/974365 | 20:18 | |
| @mnaser:matrix.org | there ya go! | 20:19 |
| @clarkb:matrix.org | +2 thanks | 20:20 |
| @fungicide:matrix.org | the held etherpad 2.6.0 test node is working fine for me, though i agree i can't find the new session transfer feature. but also i can't find it on any of the public etherpad services linked from etherpad.org either (maybe none of them have upgraded yet?) | 20:52 |
| @fungicide:matrix.org | none of them have a "wheel icon in the top right corner" of their "home page" | 20:53 |
| @clarkb:matrix.org | I really do wonder if they are doing prep work for something and this is the intended final state and it isn't actually functional yet | 20:55 |
| -@gerrit:opendev.org- Zuul merged on behalf of Mohammed Naser: [opendev/system-config] 974365: Add ProxyPreserveHost to codesearch vhost https://review.opendev.org/c/opendev/system-config/+/974365 | 21:27 | |
| @clarkb:matrix.org | mnaser: ^ that has deployed and the pop up no longer happens for me in chrome. viewing page source also shows the problem seems to have gone away. Thanks again | 21:37 |
| @clarkb:matrix.org | and now I'm going to pop out for that bike ride | 21:37 |
| @fungicide:matrix.org | i see that lists.o.o is using nearly all of its memory. there's ~10% free ram, but all 2gb of swap is showing in use which suggests there's probably spikes in memory pressure maxing it out | 22:28 |
| @fungicide:matrix.org | i'm going to restart the containers and apache to try and maybe free up some stale allocations for now | 22:28 |
| @fungicide:matrix.org | looks like the last restart was 2025-09-19, for posterity | 22:30 |
| @fungicide:matrix.org | #status log Restarted Mailman mailing list services in order to free up memory on the server | 22:35 |
| @fungicide:matrix.org | it takes a few minutes to fully start up | 22:35 |
| @fungicide:matrix.org | mmm, is statusbot on holiday? | 22:37 |
| @fungicide:matrix.org | the process is still running on eavesdrop | 22:38 |
| @fungicide:matrix.org | i have no idea how to tell if a user is in the room | 22:49 |
| @fungicide:matrix.org | element says there are 34 members here, but only shows me the names of 17 of them | 22:50 |
| @fungicide:matrix.org | i see "logs" is present but nothing i'd recognize as statusbot, so i'll give it a restart and see if it shows up | 22:50 |
| @fungicide:matrix.org | it also shows me there's a "gerrit" user in here | 22:52 |
| @fungicide:matrix.org | `2026-01-22 22:51:14,108 INFO statusbot.matrix: Join room #opendev:opendev.org` | 22:56 |
| @fungicide:matrix.org | i still don't see it listed in element | 22:56 |
| @fungicide:matrix.org | #status log Restarted Mailman mailing list services in order to free up memory on the server | 22:56 |
| @clarkb:matrix.org | fungi: the user should be status and is in this room | 23:23 |
| @fungicide:matrix.org | for some reason element doesn't show that user to me in the sidebar | 23:24 |
| @fungicide:matrix.org | it shows gerrit and logs though | 23:24 |
| @jim:acmegating.com | https://opendev.org/opendev/system-config/src/commit/3b39461cf3a369617489d3bcbef891f6392f1cb6/inventory/service/group_vars/eavesdrop.yaml#L204 | 23:24 |
| @fungicide:matrix.org | but also i don't know why it says "34 members" and then only lists the names of half that many | 23:24 |
| @jim:acmegating.com | Fungi add your name there | 23:25 |
| @fungicide:matrix.org | aha, thanks, that solves half the mystery | 23:25 |
| @jim:acmegating.com | Yeah, sorry I forgot that on Tuesday | 23:25 |
| @jim:acmegating.com | I dunno about the other half, I see all users in element | 23:26 |
| @fungicide:matrix.org | no worries | 23:26 |
| -@gerrit:opendev.org- Jeremy Stanley https://matrix.to/#/@fungicide:matrix.org proposed: [opendev/system-config] 974381: Grant fungi access to statusbot in Matrix https://review.opendev.org/c/opendev/system-config/+/974381 | 23:26 | |
| @clarkb:matrix.org | There is a button to expand the list beyond the default set iirc and a search/filter bar but I'm still on my phone which is slightly different | 23:27 |
| @jim:acmegating.com | Also, btw, these bot users will never leave the room, even if offline. They'll just get queued messages next time they connect. | 23:27 |
| @fungicide:matrix.org | yeah, in the element web client i don't have any expand or search option in the people sidebar | 23:28 |
| @fungicide:matrix.org | just an invite button, the member count, and then a list of names/avatars | 23:28 |
| @jim:acmegating.com | I'm also on phone so can't +2 but feel free to proxy my review on that change. :) | 23:29 |
| @clarkb:matrix.org | fungi: in web element you hit the room info (i) button in the top right, then hit people and it seems to show all 34 to me currently | 23:46 |
| @clarkb:matrix.org | an update must've changed element to show all people in that view because the zuul room which has a lot more people in it seems to list all of them | 23:47 |
| @clarkb:matrix.org | 974381 has been approved | 23:48 |
Generated by irclog2html.py 4.0.0 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!
