2025-01-08T01:35:41.153Z | <Dan Mick> ugh, jeez, this is cryptocode. run-tox-mgr is the name that results from a cmake call to add_tox_test(mgr,..) |
2025-01-08T01:37:37.781Z | <Dan Mick> anyway, I don't believe this has anything at all to do with the change for container/. It's long past that code, having decided that this change needs to actually be tested. |
2025-01-08T01:39:24.775Z | <Dan Mick> it's in the 'run ctest' stage of make check, and I don't know how to correlate error messages with the actual test. |
2025-01-08T01:44:51.715Z | <Dan Mick> @Christopher Hoffman what did you see that made you suspect that code at [https://github.com/ceph/ceph-build/commit/391a2970a63f3f7b798b91fdd5ebc6084b759c07#diff-1326d7a2d74cace6f00e1cd5d3[…]59eed098d2133fd72c0ec6700bR1710](https://github.com/ceph/ceph-build/commit/391a2970a63f3f7b798b91fdd5ebc6084b759c07#diff-1326d7a2d74cace6f00e1cd5d3ad15ead3fc6359eed098d2133fd72c0ec6700bR1710) ? |
2025-01-08T02:05:14.742Z | <Dan Mick> (commented on the PR; there is actually some output that seems to show a mypy error) |
2025-01-08T12:51:39.050Z | <John Mulligan> yes, the mypy error is the cause of the failure. But so far no one knows why it started failing (yet). |
2025-01-08T13:56:47.577Z | <Christopher Hoffman> Yeah- the lead I was following was a red herring, the Jenkins highlight was misleading. Thanks to you both for taking a look
It looks like lead I was looking down was a red herring. Thanks |
2025-01-08T15:49:19.907Z | <Christopher Hoffman> Yeah- the lead I was following was a red herring, the Jenkins highlight was misleading. Thanks to you both for taking a look |
2025-01-08T16:25:51.630Z | <Casey Bodley> weekly rgw meeting soon in [ <https://pad.ceph.com/p/rgw-weekly](https://meet.google.com/mmj-uzzv-qce> ) |
2025-01-08T16:27:14.090Z | <Yehuda Sadeh-Weinraub> @Casey Bodley looking at `int RGWBucketInstanceMetadataHandler::put_prepare( )`, any reason we guard layout only in the `from_remote_zone` case? Seems a bit convoluted |
2025-01-08T16:28:19.113Z | <Casey Bodley> my guess is that we wanted `radosgw-admin metadata put` to be able to modify local layout info |
2025-01-08T16:30:42.569Z | <Yehuda Sadeh-Weinraub> Maybe? In any case it's probably better to make some kind of explicit flag to allow local changes whether it's from remote or not, and should probably prevent overriding it for local opts too by default |
2025-01-08T16:30:58.575Z | <Yehuda Sadeh-Weinraub> Maybe? In any case it's probably better to make some kind of explicit flag to allow local changes whether it's from remote or not, and should probably prevent overriding it for local ops too by default |
2025-01-08T16:55:16.421Z | <Casey Bodley> we wouldn't handle abitrary layout changes correctly (like num shards from 11 -> 20) anyway |
2025-01-08T16:56:04.606Z | <Casey Bodley> that _might_ be the only way to enable indexless buckets atm, but that wouldn't delete existing index objects either |
2025-01-08T16:58:28.317Z | <Yehuda Sadeh-Weinraub> it's be better to make it all explicit |