opendevreview | Merged openstack/releases master: Add releasenotes link for Tempest 2024.1 release https://review.opendev.org/c/openstack/releases/+/914822 | 09:20 |
---|---|---|
opendevreview | Merged openstack/releases master: Release keystone for stable/zed https://review.opendev.org/c/openstack/releases/+/906848 | 09:20 |
opendevreview | Merged openstack/releases master: [telemetry-tempest-plugin] Tag yoga-last https://review.opendev.org/c/openstack/releases/+/908365 | 09:20 |
noonedeadpunk | hey folks! I've just spotted that ussuri-em tag is missing somehow for openstack-ansible repo? While I would expect to see it according to https://opendev.org/openstack/releases/src/branch/master/deliverables/ussuri/openstack-ansible.yaml#L92-L95 | 10:43 |
frickler | noonedeadpunk: it seems the tag is in place, I also see it in my local repo https://opendev.org/openstack/openstack-ansible/src/tag/ussuri-em , but it does seem to be missing in the tag dropdown, might be a gitea bug? | 10:54 |
noonedeadpunk | yeah, actually you're right here, I indeed do see that in local repo | 10:56 |
noonedeadpunk | and it's visible in github | 10:57 |
noonedeadpunk | so indeed just related to gitea somehow | 10:57 |
noonedeadpunk | whatever then :D | 10:57 |
noonedeadpunk | sorry, I in fact did not double-checked on that before writing, as did not expect it could be really a gitea thing | 10:58 |
frickler | fwiw I'm seeing the same thing for kolla, so no special treatment for osa ;) but only that particular tag is missing, I wonder what can be special about it | 11:08 |
noonedeadpunk | I guess for all projects actually | 11:43 |
noonedeadpunk | but yes, what is special about tag is a great question | 11:43 |
*** sfinucan is now known as stephenfin | 12:25 | |
fungi | we could test forcing replication and see if it was somehow missed previously | 13:58 |
fungi | the bulk tagging and branching operations flood gerrit with replication events, so it's possible there's a hole some of them might have fallen into | 13:58 |
fungi | maybe it overwhelms gitea? | 13:59 |
fungi | youtube lgtm | 14:00 |
fungi | er, wrong channel | 14:00 |
frickler | fungi: well gitea does seem to show the tag if you use the URL I showed above, so I don't think it is an replication issue? doesn't gitea use a local db though for some stuff? maybe I should check first whether the results are the same for each gitea instance | 14:23 |
fungi | the "tags" list is tied to its releases feature, so there might be something specific to that | 14:28 |
fungi | and yeah, that's probably recorded in the rdbms rather than generated directly from the repo | 14:29 |
fungi | in which case re-replicating may not help | 14:30 |
clarkb | yes this is a known issue. Not sure why it happens yet. The content is in git just fine so rereplicating is a noop | 14:41 |
clarkb | the problem seems to be that gitea isn't recording the tag(s) in its database when the tag is pushed. And the same tag will be present in some repos and not in others | 14:41 |
clarkb | after finding the case sensitivity bug I suspect its the same area of code that is having problems (but in this situation case sensitivity doesn't seem to tbe the problem) | 14:41 |
frickler | I was just looking at the gitea db and I found a "branch" table, which has the expected entries, but none for the tags | 14:45 |
clarkb | that branch table is what has the case sensitivity problems. I wonder where they load tags from | 14:45 |
fungi | more caracal media coverage: | 18:42 |
fungi | https://www.theregister.com/2024/04/04/openstack_caracal_released/ | 18:43 |
fungi | https://techcrunch.com/2024/04/04/openstack-improves-support-for-ai-workloads/ | 18:43 |
Generated by irclog2html.py 2.17.3 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!