| fungi | i can see one of the reasons the openeuler mirror seems to grow without bound. they never remove older packages, so there are currently 10 different versions of firefox for this one release we're mirroring | 00:18 |
|---|---|---|
| Clark[m] | Centos does the same thing | 01:11 |
| opendevreview | Dr. Jens Harbott proposed openstack/project-config master: Revert "Move test-release-openstack job to Noble" https://review.opendev.org/c/openstack/project-config/+/937085 | 05:13 |
| fungi | openeuler mirror refresh finally completed at 06:37 utc. i ran another manual pass just now to make sure it was effectively a no-op and it completed in under a minute. i'm releasing the lock i've been holding on it now and closing the screen session | 13:17 |
| fungi | quota usage is under 70% for the moment, but we also know we can probably drop it quite a bit if we exclude the various installer/boot image directories and riscv architecture subtrees | 13:19 |
| opendevreview | Merged openstack/project-config master: Revert "Move test-release-openstack job to Noble" https://review.opendev.org/c/openstack/project-config/+/937085 | 13:23 |
| *** ykarel__ is now known as ykarel | 13:26 | |
| opendevreview | Pierre Riteau proposed openstack/project-config master: Fix release name in Launchpad comment https://review.opendev.org/c/openstack/project-config/+/937184 | 14:11 |
| opendevreview | Merged openstack/project-config master: Fix release name in Launchpad comment https://review.opendev.org/c/openstack/project-config/+/937184 | 14:23 |
| clarkb | fungi: ++ to excluding things we know we won't need in the openeuler mirror | 16:03 |
| frickler | was there some change in openeuler mirroring? I guess I missed that | 16:05 |
| fungi | frickler: yeah, https://review.opendev.org/c/opendev/system-config/+/927462 | 16:07 |
| clarkb | upstream did merge my openid redirect fix to stable-3.9. We could rebuild images and redeploy but the change hasn't landed on 3.10 yet | 16:15 |
| clarkb | I think a better approach is to shelve fixing that in our environment until after the 3.10 upgrade then get the change moved to 3.10 in gerrit | 16:15 |
| clarkb | and do a quick image update post upgrade sometime to pull in the fix | 16:15 |
| fungi | sgtm | 16:15 |
| fungi | thanks for working through that! | 16:15 |
| fungi | gonna disappear to grab a quick bite, but can put together the openeuler exclusion list when i get back in a bit | 16:22 |
| clarkb | sounds good thnaks. I am having an exceptionally slow start today. Side effect of getting all the things done yesterday and the really cold morning today I think | 16:23 |
| opendevreview | Jay Faulkner proposed openstack/diskimage-builder master: Fix pyyaml install for svc-map https://review.opendev.org/c/openstack/diskimage-builder/+/937192 | 16:33 |
| opendevreview | Clark Boylan proposed openstack/diskimage-builder master: Disable openeuler mirror use in dib-functests https://review.opendev.org/c/openstack/diskimage-builder/+/937200 | 18:39 |
| tonyb | I don't if it's been noted here but Fedora is about to migrate to ForgeIO https://fedoramagazine.org/fedora-moves-towards-forgejo-a-unified-decision/ | 19:13 |
| tonyb | I haven't read the history but gitea doesn't appear to have been considered ? | 19:14 |
| clarkb | I think the whole gitea forgejo split created a lot of strong opinions and divisions | 19:15 |
| fungi | and fud | 19:15 |
| clarkb | I didn't manage to follow it very closely but I think both sides made mistakes and unfortunately this led to a lack of interest in trying to find common ground | 19:16 |
| clarkb | and now I'm mostly just trying to do what is the least amount of work for us | 19:17 |
| clarkb | functionally they seem very similar still particularly since we are not big feature users | 19:17 |
| tonyb | That's fair. | 19:18 |
| clarkb | I think forgejo still accuses gitea of requiring copyright assignment which as far as I can tell (I'm not a lawyer) has never been the case and continues to not be the case | 19:19 |
| clarkb | instead they use a policy very similar to kubernetes where they don't put individual copyright statements in every file to avoid files growing large with copyright statements | 19:20 |
| clarkb | instead refering to copyright as being owned by the project authors and then you refer to the git record to know who owns what | 19:20 |
| corvus | whether you put a statement in a file is orthogonal to who actually holds the copyright | 19:20 |
| fungi | precisely | 19:21 |
| clarkb | and unfortunately those discussions very quickly grew out of control due to other disagreements between the groups and ya | 19:22 |
| fungi | adding "Copyright Gitea Contributors" in files and refusing addition of other copyright headers isn't "copyright assignment" by any definition i'm aware of | 19:22 |
| fungi | from what i saw at the time, there were accusations by someone who had collected third-party funding for specific gitea feature work that the money was mis-spent, and that person too the opportunity of the gitea maintainers forming a corporate entity for holding the trademark to come up with lots of doom-n-gloom predictions which still haven't come to pass, but convinced others to go | 19:25 |
| fungi | in on a fork instead | 19:25 |
| fungi | and now they can't take back what they've claimed without looking foolish, so have continued to double-down on baseless accusations instead | 19:26 |
| clarkb | I think some of the criticisms are valid fwiw | 19:26 |
| clarkb | the security repsonse and reporting process has been quite lackluster | 19:26 |
| fungi | yes, which makes it hard to dismiss the others | 19:27 |
| tonyb | Well that's all a little demoralising, I'm a little sorry I brought it up | 19:29 |
| clarkb | ya as a largely third party observer the unfortunate thing is that if some common ground could've been found then likely we would've had a better project overall instead we have two separate projects that don't collaboarate | 19:29 |
| clarkb | because there were definitely areas for improvement and interest from the community | 19:30 |
| clarkb | I should also say there are areas for improvement | 19:31 |
| opendevreview | Jeremy Stanley proposed opendev/system-config master: Trim unneeded content from OpenEuler mirror https://review.opendev.org/c/opendev/system-config/+/937208 | 20:48 |
| fungi | i tested ^ manually from mirror-update with the lock held, and it deleted the stuff i had noticed at least. it will still be a few minutes before https://grafana.opendev.org/d/9871b26303/afs?orgId=1&viewPanel=44 reflects the decrease in utilization | 20:49 |
| clarkb | I went ahead and approved it since it has already been used and if we resync without it we'll redownload stuff we don't want | 20:56 |
| fungi | thanks, i've still got the lock held but will drop it once that deploys | 20:59 |
| fungi | that removed 60gb of data from the volume, reducing utilization from 69% of quota to 51% | 21:48 |
| fungi | i suppose we could consider lowering the quota again (it's at 350gb right now), but given the tendency for this data to grow without bound maybe it's better if we don't | 21:49 |
| opendevreview | Merged opendev/system-config master: Trim unneeded content from OpenEuler mirror https://review.opendev.org/c/opendev/system-config/+/937208 | 22:00 |
| clarkb | fungi: before we call it a day today is there anything else we need or want to do before the gerrit upgrade tomorrow? As mentioned yesterday I think I went through a reaosnable amount of testing and checking things without overdoing it | 22:19 |
| fungi | clarkb: nothing comes to mind, i think we're all set. i've got reminders set up for timing the advance steps | 22:59 |
| clarkb | cool and my morning alarm should have me up more than early enough | 22:59 |
| fungi | since the deploy jobs for 937208 completed successfully, i've dropped the lock i was holding | 23:05 |
| clarkb | looking at the fedora evaluation of forgejo one interesting thing is being able to depends-on between issues and not just code | 23:34 |
| clarkb | but also it looks like they are giving up on gating with forgejo? | 23:35 |
| clarkb | or maybe they are just not setting it up for individual package maintainers? either way that does seem like a big regression but I don't maintain packages so maybe not | 23:36 |
Generated by irclog2html.py 2.17.3 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!