[03:28] <mwhudson> wgrant: i'm trying to test some apport integration by reporting bugs against staging but it seems the bug data on staging is never processed?
[03:29] <mwhudson> wgrant: e.g. https://bugs.staging.launchpad.net/subiquity/+filebug/9e6444e6-c3c3-11e9-9244-18a9056479cc?field.title=block+probing+failed+with+ZeroDivisionError has been refreshing every 10s for several minutes now
[03:29] <mup> Bug #9: Rosetta's po parser is too strict <lp-translations> <Launchpad itself:Fix Released by carlos> <https://launchpad.net/bugs/9>
[03:29] <mwhudson> is this expected or did i break it somehow?
[03:31] <wgrant> mwhudson: Hm, the processing job runs minutely on staging
[03:31] <wgrant> It's possible it's broken somehow
[03:33] <mwhudson> wgrant: is that something you can determine or do i need a webop?
[03:35] <wgrant> mwhudson: I'm having a look
[03:35] <mwhudson> wgrant: tia
[03:35] <wgrant> But logrotation is apparently misconfigured so the log is 3.8GB
[03:35] <wgrant> will take a while to sync...
[03:35] <mwhudson> oof
[03:55] <wgrant> mwhudson: Well that's weird. It got a 503 while trying to talk to the librarian.
[03:55] <wgrant> Have you tried more than once?
[03:55] <wgrant> Some caches may have been cold or something
[03:55] <mwhudson> wgrant: i've tried twice
[03:55] <mwhudson> and uhh
[03:56] <mwhudson> wgrant: so it might be worth trying a few times?
[03:56] <wgrant> mwhudson: The answer to that on staging is always yes
[03:56] <wgrant> It's still got the full terabyte database, but the hot set is rarely hot
[03:58]  * mwhudson tries three more times
[04:03] <mwhudson> wgrant: my other three reports don't seem to be being processed either :(
[04:04] <wgrant> Hmm. Let's get IS to investigate after this network maintenance, then.
[04:08] <mwhudson> ok
[04:08] <mwhudson> well i need to go actually, so i'll try again tomorrow and then bug a sysadmin if it still doesn't work
[09:09] <cjwatson> SpecialK|Canon: Moving to Jenkins is a possibility, but it'd require a fair bit of work to make sure we don't seriously regress test runtime (would probably need us to run on lots of parallel instances or something).  What I was actually planning to do when we finish the move to git was to use Jenkins to process approved merge requests but without running tests, so it would replace bzr lp-land ...
[09:09] <cjwatson> ... rather than replacing buildbot.
[09:46] <wgrant> I mean the current parallel buildbot slaves are on super beefy machines from 2011
[09:46] <wgrant> Which means we could probably get reasonable performance from some moderately beefy VMs from 2019
[09:46] <cjwatson> Right, I'm sure it can be done, but it probably wouldn't work well if done naively.
[09:47] <cjwatson> Needs some attention rather than just whacking stuff into ols-jenkaas.
[10:32] <wgrant> Yep