[08:48] waveform: added you to ~motu and announced. sil2100, hopefully I've covered both the things, let me know if I'm missing anything. TIA! \o/ [12:05] checking in for day 2 of +1 maintenance [12:10] hi, is there a way to avoid the "File has unexpected size […]. Mirror sync in progress" errors on our Ubuntu mirrors? we only update them twice a day and apparently we sometimes sync them when the source rsync server is in a wrong state (or files changed during the rsync process, i'm not sure) [12:13] ok from the logs i see that i started rsyncing, then the Release file changed during my sync :/ [12:21] I'm pretty sure it's possible to reliably sync without errors, but I'm not up to date with the tooling that makes that easy [12:26] rbasak: the best idea i have right now would be to repeatedly rsync until it finds nothing to sync… but i don't know how long that would take [12:29] > Do not use your own scripts, and do not just use single-pass rsyncs [12:29] https://www.debian.org/mirror/ftpmirror [12:29] i guess i'm just not using the right tool [12:31] sbraz: yeah if you just rsync then you'll be reporting the availability of new debs to your consumers before they've actually arrived, and your consumers will get apt errors. [12:32] rbasak: yeah makes total sense, i guess we're just lucky that we don't hit this very often, i'll read https://salsa.debian.org/mirror-team/archvsync/ which has more info about 2-stage sync and the files to exclude from the first stage [13:58] LocutusOfBorg: hi, did you understand why cmake is failing dep8 tests? I ran them in a lunar container with autopkgtest and I don't get the same badpkg, I don't quite understand what is going on [13:58] but seems related to curl [13:59] maybe re-run with all-proposed? But it could be days before we see a result [14:00] or maybe re-run with just curl from proposed [14:04] * ahasenack reruns with the exact same --apt-pocket cmdline [14:05] ahasenack, already retried some hours ago [14:05] it was due to perl [14:05] I guess [14:06] I saw your retries [14:06] I'm testing locally [14:06] I uploaded perl rebuilds before openings, but sadly cmake was uploaded before it [14:06] do you have other retries queued up? [14:06] so, yes they need a retry and probably -proposed pocket is not needed anymore now [14:06] I already retried some perl and node stuff, but until queues settles down, I don't see much value in truing to find what needs a retry [14:06] https://autopkgtest.ubuntu.com/packages/c/cmake/lunar/amd64 shows two failed ones [14:06] I can't even load the excuses page [14:07] (two failed retries from you, I mean) [14:08] so perl failed to use proposed perl [14:08] :/ [14:08] retried [14:13] rbasak: i can't find much doc about using ftpsync with ubuntu, i see the doc mentions debmirror, do you know anyone here who could answer as to what is the preferred method for ubuntu mirroring? i assume ftpsync works but i might as well use what every official mirror uses [14:16] LocutusOfBorg: this just work on a fresh lunar container (the target system): autopkgtest -o dep8 -B -U -s --apt-pocket=proposed=src:mercurial,src:check-manifest,src:cmake,src:composer,src:devpi-common,src:setuptools-scm cmake -- ssh -H 10.109.225.205 -r --capability isolation-container --capability revert-full-system [14:17] I see your last failed retry used all-proposed [14:17] one of the other failed ones used the cmdline I pasted above (for --apt-pocket) === popey9 is now known as popey [14:22] sbraz: I'm aware of debmirror and apt-mirror but that's the limit of my knowledge, sorry. [14:22] Something that's aware of by-hash would be ideal. [14:22] But I don't know to what extent those tools have picked that up. [14:23] https://wiki.ubuntu.com/AptByHash [14:23] sbraz, did you consider asking in #ubuntu-mirrors how others prevent this ? [14:23] Apparently ubumirror exists as well [14:24] heh ... well, ubuntu-mirrors is linked from https://wiki.ubuntu.com/Mirrors [14:35] ogra: i wasn't aware of this chan, i'll join it, thanks [14:36] that wiki page links to https://wiki.ubuntu.com/Mirrors/Scripts which contains interesting tools as well [14:37] yep 🙂 [14:37] (and unlike 99.9% of the ubuntu wiki pages, this one seems to actualy be well up to date) === popey0 is now known as popey === ben is now known as Guest1083 === ben__ is now known as ben === popey1 is now known as popey