| opendevreview | Matthew Oliver proposed openstack/swift master: trace: Add RequestTraceMiddleware with OpTel https://review.opendev.org/c/openstack/swift/+/857559 | 08:15 |
|---|---|---|
| opendevreview | Merged openstack/swift master: sharder: merge shard shard_ranges from root while sharding https://review.opendev.org/c/openstack/swift/+/852905 | 10:07 |
| zigo | I have a very quick ring and weight question. | 14:15 |
| zigo | Currently, we have many clusters with weight 100, 133, 150, etc., and that's very annoying. I wish to move to a "weight = size of the disk" type of thing. | 14:15 |
| zigo | If I do that, with some replace_weight thingy, is that ok? Swift wont shuffle all partitions when doing the rebalance? | 14:15 |
| zigo | timburke: ^ | 14:20 |
| zigo | I just tried, I think that's kind of fine to do it ! :) | 14:30 |
| timburke | zigo, yes, should be fine to do! as long as all disks are scaled by the same (non-zero) factor, nothing should move | 16:35 |
| timburke | 🤔 might be nice if swift-ring-builder had some scale_weights command to do that for you... | 16:36 |
| timburke | oh yeah! zigo, i remember you mentioning some failing unit tests on 32bit arches -- did that get resolved? https://qa.debian.org/excuses.php?package=swift is looking better now. iirc it looked like off_t maybe got resized... | 16:45 |
| zigo | timburke: It did migrate because I have blacklisted the failing unit tests. | 19:50 |
| zigo | # These are failing on 32 bits: | 19:50 |
| zigo | # test.unit.common.test_utils.TestPunchHoleReally.test_punch_a_hole | 19:50 |
| zigo | # test.unit.common.test_utils.Test_LibcWrapper.test_argument_plumbing | 19:50 |
| zigo | That's what I have in debian/tests/unittests currently. | 19:50 |
| timburke | zigo, got it, thanks. i'll see if i can find a way to reliably get an off_t from python... | 20:30 |
| timburke | maybe ssize_t would be close enough? not sure how best to go about finding which (if any) platforms have those not the same size... | 20:45 |
| timburke | or maybe long_t ? | 20:51 |
| timburke | :-/ but then, fallocate won't work for uploads over 2GB on those 32 bit platforms... we could switch the test to use lseek64 easily enough, but i'm not sure there's a 64-bit equivalent of fallocate on 32 bit platforms... | 21:00 |
| timburke | huh. https://github.com/python/cpython/blob/v3.11.0/Lib/ctypes/__init__.py#L483-L491 looks sus to me | 21:10 |
| opendevreview | Ghanshyam proposed openstack/swift master: Temporary pin the swift-dsvm-functional nodeset to Focal https://review.opendev.org/c/openstack/swift/+/864946 | 21:30 |
| gmann | timburke: ^^ pining nodeset to Focal to allow devstack base job migration patch to merge tomorrow. you can revert it issue it is fixed. appreciate if you can merge it before tomorrow. thanks | 21:32 |
| opendevreview | ASHWIN A NAIR proposed openstack/swift master: proxy-server exception logging shows replication_ip/port https://review.opendev.org/c/openstack/swift/+/860866 | 23:39 |
Generated by irclog2html.py 2.17.3 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!