| zigo | Can someone help me figuring out why Swift segfault under Python 3.11 ? | 10:38 |
|---|---|---|
| zigo | I don't feel I can easily file a bug report because I don't have much details ... | 10:39 |
| zigo | timburke: ^ | 11:34 |
| opendevreview | Merged openstack/swift master: proxy: refactor ContainerController._GET_using_cache https://review.opendev.org/c/openstack/swift/+/865706 | 13:09 |
| opendevreview | Merged openstack/swift master: proxy: add test for ContainerController._GET_using_cache https://review.opendev.org/c/openstack/swift/+/865707 | 17:44 |
| timburke | zigo, seems like an issue with the sqlite module -- i can get a similar segfault pretty easily. running under pytest, i at least get a python call stack: https://paste.opendev.org/show/bV9t8Ujs6Es7y2sowPgy/ | 18:22 |
| timburke | which points to a pretty boring bit in swift's tests: https://github.com/openstack/swift/blob/2.30.0/test/unit/common/test_db.py#L615 | 18:22 |
| timburke | the next bit in the stack seems to be resolving the context manager at https://github.com/openstack/swift/blob/2.30.0/swift/common/db.py#L540-L566 | 18:46 |
| opendevreview | Alistair Coles proposed openstack/swift master: WIP: restructure cached updating shard ranges https://review.opendev.org/c/openstack/swift/+/863562 | 18:49 |
| opendevreview | Alistair Coles proposed openstack/swift master: WIP: restructure cached listing shard ranges https://review.opendev.org/c/openstack/swift/+/865708 | 18:49 |
| timburke | i'm still not clear on why we don't get a frame pointing at some bit of that code, though... maybe we've already gotten through it all and dropped back to cpython to raise the StopIteration? | 18:56 |
| timburke | anyway, adding some print debugging while still *within* the context manager yielded some interesting results: i can print immediately after the assignment to `rows`, i can do `rows[0]`, i can do `rows[0]['account']`, but if i try to print all of `rows[0]` or do something that would trip a KeyError, segfault! | 19:04 |
| timburke | so maybe it's actually something in the dict implementation? | 19:06 |
| opendevreview | ASHWIN A NAIR proposed openstack/swift master: Malformed CompleteMultipartUpload request should 400 https://review.opendev.org/c/openstack/swift/+/865348 | 19:10 |
| timburke | zigo, fwiw all my core dumps pop in pymalloc_alloc, at https://github.com/python/cpython/blob/v3.11.0/Objects/obmalloc.c#L1979 | 20:06 |
| timburke | nbytes tends to be small, in the 80-100 range | 20:07 |
| timburke | this one seemed particularly odd to me: https://paste.opendev.org/show/bAlT69FH4MYXwgFZIz6n/ -- it seems to have tripped in the process of raising that StopIteration... | 20:13 |
| opendevreview | Matthew Oliver proposed openstack/swift master: WIP sharder: update own_sr stats explicitly https://review.opendev.org/c/openstack/swift/+/852283 | 22:56 |
| opendevreview | Matthew Oliver proposed openstack/swift master: Sharding: No stat updates before CLEAVED state https://review.opendev.org/c/openstack/swift/+/837811 | 22:57 |
| opendevreview | Matthew Oliver proposed openstack/swift master: WIP: sharding: Block moving to CLEAVED if cleaved rows < expected https://review.opendev.org/c/openstack/swift/+/843973 | 22:57 |
Generated by irclog2html.py 2.17.3 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!