| *** mattw4 has quit IRC | 00:06 | |
| *** igordc has quit IRC | 00:28 | |
| *** jamesmcarthur has joined #zuul | 00:34 | |
| *** igordc has joined #zuul | 00:41 | |
| *** jamesmcarthur has quit IRC | 00:46 | |
| *** michael-beaver has quit IRC | 00:55 | |
| *** jamesmcarthur has joined #zuul | 01:11 | |
| *** jamesmcarthur has quit IRC | 01:35 | |
| *** logan- has quit IRC | 01:43 | |
| *** logan- has joined #zuul | 01:46 | |
| *** igordc has quit IRC | 01:50 | |
| *** mhu has quit IRC | 02:11 | |
| *** jamesmcarthur has joined #zuul | 02:34 | |
| *** rfolco|ruck has quit IRC | 03:06 | |
| *** michael-beaver has joined #zuul | 03:53 | |
| *** jamesmcarthur has quit IRC | 04:00 | |
| *** michael-beaver has quit IRC | 06:03 | |
| *** gtema has joined #zuul | 08:21 | |
| *** bjackman has joined #zuul | 08:34 | |
| *** gtema has quit IRC | 08:52 | |
| *** bjackman has quit IRC | 08:53 | |
| *** zxiiro has quit IRC | 09:44 | |
| *** mnaser has quit IRC | 09:44 | |
| *** guilhermesp has quit IRC | 09:44 | |
| *** iamweswilson has quit IRC | 09:44 | |
| *** clayg has quit IRC | 09:44 | |
| *** clayg has joined #zuul | 09:47 | |
| *** guilhermesp has joined #zuul | 09:47 | |
| *** zxiiro has joined #zuul | 09:48 | |
| *** mnaser has joined #zuul | 09:59 | |
| *** iamweswilson has joined #zuul | 10:00 | |
| *** bjackman has joined #zuul | 10:33 | |
| *** bjackman has quit IRC | 10:44 | |
| *** bjackman_ has joined #zuul | 10:44 | |
| *** bjackman_ has quit IRC | 10:48 | |
| *** bjackman has joined #zuul | 10:48 | |
| *** tosky has joined #zuul | 11:20 | |
| *** bjackman has quit IRC | 11:44 | |
| *** jhesketh has quit IRC | 14:06 | |
| *** altlogbot_0 has quit IRC | 14:48 | |
| *** altlogbot_1 has joined #zuul | 14:53 | |
| *** clarkb has quit IRC | 15:23 | |
| *** paladox has quit IRC | 15:24 | |
| *** paladox has joined #zuul | 15:28 | |
| *** clarkb has joined #zuul | 15:37 | |
| *** tosky has quit IRC | 15:43 | |
| *** paladox has quit IRC | 15:54 | |
| *** paladox has joined #zuul | 15:54 | |
| *** jamesmcarthur has joined #zuul | 16:01 | |
| *** jamesmcarthur has quit IRC | 16:06 | |
| *** yolanda has quit IRC | 17:22 | |
| fungi | got a weird quandry... | 18:00 |
|---|---|---|
| fungi | in opendev we build debian-derivative distro package mirrors with reprepro, and because it generates its own package indices based on what packages it mirrored for a repository it doesn't generate a repository when the one it's been asked to mirror has no packages (that is to say the official mirrors are present but have empty indices, and so our mirror just doesn't get created at all) | 18:01 |
| fungi | this is presently happening with our debian buster-updates mirror. normally buster-updates would hold the files for debian stable point releases of buster, but because buster only released very recently there have been none yet | 18:02 |
| fungi | i see that in the configure-mirrors role in zuul-jobs we unconditionally include the updates.list template: https://opendev.org/zuul/zuul-jobs/src/branch/master/roles/configure-mirrors/tasks/mirror/Debian.yaml#L11 | 18:03 |
| fungi | i'm trying to figure out what the cleanest way is to omit it for debian-buster images but not debian-stretch images | 18:04 |
| fungi | (i already exhausted possible options for forcing reprepro to create empty repository indices) | 18:04 |
| fungi | should that line just be wrapped in a conditional for a new role variable we can use to toggle it off for certain images? | 18:08 |
| fungi | or turn that whole block into a parameterized list with supplied defaults? | 18:09 |
| SpamapS | fungi: you could make the conditional based on a check whether or not the source repo is empty. | 19:18 |
| SpamapS | when: repos[item].package_count > 0 | 19:19 |
| SpamapS | something like that | 19:19 |
| fungi | hrm, except the repository isn't there on our mirrors at all because of reprepro not generating indices for packageless repositories | 19:19 |
| fungi | so apt update breaks | 19:19 |
| fungi | so it would need to be conditional on whether the official debian copy of buster-updates is still packageless | 19:20 |
| fungi | ideally there would just be an option to force reprepro to create empty indices for a repository when there were no packages to mirror, but i've thus far been unable to figure out how to do that if it's even possible | 19:21 |
| fungi | i guess it could be made conditional on attempting to fetch the url | 19:22 |
| fungi | so if this is a foo image and your mirror lacks foo-updates then don't include that template | 19:23 |
| *** tosky has joined #zuul | 19:25 | |
| *** pcaruana has joined #zuul | 19:30 | |
| *** jamesmcarthur has joined #zuul | 19:46 | |
| *** jamesmcarthur has joined #zuul | 19:46 | |
| *** pcaruana has quit IRC | 20:14 | |
| *** jamesmcarthur has quit IRC | 20:51 | |
| *** jamesmcarthur has joined #zuul | 20:55 | |
| *** jamesmcarthur has quit IRC | 21:34 | |
| *** rfolco|ruck has joined #zuul | 22:23 | |
| *** jamesmcarthur has joined #zuul | 22:34 | |
| *** jamesmcarthur has quit IRC | 22:37 | |
| *** jamesmcarthur has joined #zuul | 22:41 | |
| *** saneax has joined #zuul | 22:44 | |
| *** jamesmcarthur has quit IRC | 22:57 | |
| corvus | fungi: i *think* the new mirror_info structure i started on can accomodate that. we haven't implemented a new version of config-mirrors based on it yet, but maybe that should be the first one. https://zuul-ci.org/docs/zuul-jobs/mirror.html | 23:16 |
| corvus | fungi: if you have a sec to take a look at that data structure and see if it would let you express what you want, and if not, what's missing, that would be great | 23:16 |
| *** saneax has quit IRC | 23:16 | |
| corvus | (ie, maybe the 'components' attr there would cover it, or would we need something else?) | 23:17 |
| *** tosky has quit IRC | 23:27 | |
| fungi | oh! right, i forgot you'd redesigned that | 23:33 |
| fungi | i'll try and take a look tonight or tomorrow, that may be just the ticket | 23:33 |
| *** jamesmcarthur has joined #zuul | 23:53 | |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!