| seongsoocho | I'm trying to use the dark data delete feature in object auditor. The phenomenon I experienced while using it is that it keeps replicating on other nodes with dark data by the replicator, making the dark data immortal. Do I need to stop the replicator to do it with action=delete in auditor? | 01:13 |
|---|---|---|
| opendevreview | Yan Xiao proposed openstack/swift master: labeled-metrics: Alternative approach to emit legacy and labeled metrics https://review.opendev.org/c/openstack/swift/+/908244 | 03:53 |
| timburke | seongsoocho, very likely, yes. or, run it in logging mode and use the logs to inform some other process to do the deletion (either by unlinking on each primary at roughly the same time, or issuing a DELETE and laying down a tombstone) | 04:21 |
| seongsoocho | timburke: a-ha.. Using the logs seemd like better option. | 04:25 |
| opendevreview | Tim Burke proposed openstack/swift master: Be willing to send hashes as JSON https://review.opendev.org/c/openstack/swift/+/758638 | 05:30 |
| opendevreview | Matthew Oliver proposed openstack/swift master: internal_client: Add iter_{shard_ranges,namespaces,namespaces_from_cache} interfaces https://review.opendev.org/c/openstack/swift/+/877584 | 07:10 |
| opendevreview | Tim Burke proposed openstack/swift master: replicator: Add sync_batches_per_revert option https://review.opendev.org/c/openstack/swift/+/839649 | 07:16 |
| opendevreview | Tim Burke proposed openstack/swift master: replicator: Rename update_deleted to revert https://review.opendev.org/c/openstack/swift/+/908252 | 07:16 |
| opendevreview | Tim Burke proposed openstack/swift master: diskfile: Ignore invalid suffixes in invalidations file https://review.opendev.org/c/openstack/swift/+/908253 | 07:16 |
| opendevreview | Tim Burke proposed openstack/swift master: diskfile: Add API to remove a suffix from hashes.pkl https://review.opendev.org/c/openstack/swift/+/908254 | 07:16 |
| opendevreview | Alistair Coles proposed openstack/swift master: proxy-server: de-duplicate _get_next_response_part method https://review.opendev.org/c/openstack/swift/+/890919 | 11:08 |
| opendevreview | Alistair Coles proposed openstack/swift master: proxy: don't use recoverable_node_timeout with x-newest https://review.opendev.org/c/openstack/swift/+/908287 | 11:08 |
| opendevreview | Alistair Coles proposed openstack/swift master: CooperativeIterator: ignore period<=0 https://review.opendev.org/c/openstack/swift/+/908300 | 12:16 |
| zigo | timburke: clarkb: My tool needs to check the available space in partitions in /srv/node, so I was wondering if in the functional test suite of Swift, if we have this already provisioned, somehow, and if yes, in which path. | 12:34 |
| clarkb | zigo: the swift/tools/test-setup.sh script is what the upstream ci jobs for python unittests run to configure disk stuff | 16:06 |
| timburke | zigo, i wouldn't worry about it for a functional test; unit test (with mocked out return values for free space) should be sufficient | 16:16 |
| zigo | timburke: I'll try, thanks. | 16:17 |
| opendevreview | Tim Burke proposed openstack/swift master: Add ClosingIterator class; be more explicit about closes https://review.opendev.org/c/openstack/swift/+/904652 | 17:26 |
| opendevreview | Alistair Coles proposed openstack/swift master: DNM: WIP: test jig for native mpu audit https://review.opendev.org/c/openstack/swift/+/908337 | 19:08 |
| timburke | #startmeeting swift | 21:00 |
| opendevmeet | Meeting started Wed Feb 7 21:00:12 2024 UTC and is due to finish in 60 minutes. The chair is timburke. Information about MeetBot at http://wiki.debian.org/MeetBot. | 21:00 |
| opendevmeet | Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. | 21:00 |
| opendevmeet | The meeting name has been set to 'swift' | 21:00 |
| timburke | who's here for the swift meeting? | 21:00 |
| mattoliver | o/ | 21:00 |
| acoles | o/ | 21:00 |
| mattoliver | I got the time right today! | 21:00 |
| timburke | 🎉 | 21:01 |
| timburke | i neglected to update the agenda, but we can start with updates on last week's items | 21:02 |
| timburke | #topic swiftclient reviews | 21:02 |
| timburke | acoles reviewed p 904652; sounds like yan will be addressing comments later this week | 21:03 |
| patch-bot | https://review.opendev.org/c/openstack/swift/+/904652 - swift - Add ClosingIterator class; be more explicit about ... - 7 patch sets | 21:03 |
| timburke | i didn't get to poking at the path-normalization bug, unfortunately | 21:03 |
| timburke | clayg helped characterize the issue in https://review.opendev.org/c/openstack/python-swiftclient/+/907118 | 21:04 |
| patch-bot | patch 907118 - python-swiftclient - WIP: test for invalid segment name - 3 patch sets | 21:04 |
| timburke | #topic s3 compat | 21:04 |
| timburke | the part number patches are still in review, but seem likely to merge soon-ish | 21:05 |
| timburke | on the follow-ups, i think https://review.opendev.org/c/openstack/swift/+/906906 should probably get squashed into the main s3api patch | 21:05 |
| patch-bot | patch 906906 - swift - s3api: return 400 if partNumber and Range sent - 3 patch sets | 21:05 |
| timburke | while https://review.opendev.org/c/openstack/swift/+/906391 can probably continue to be a follow-up | 21:06 |
| patch-bot | patch 906391 - swift - slo: Support range requests for part-number queries - 2 patch sets | 21:06 |
| acoles | sounds like a plan | 21:06 |
| timburke | acoles also wrote up some more MPU cross-compat tests: https://review.opendev.org/c/openstack/swift/+/907751 | 21:06 |
| patch-bot | patch 907751 - swift - WIP s3api: add more MPU cross-compat tests - 1 patch set | 21:06 |
| mattoliver | kk, I promise to get to review these myself over the next week. They are on my list.. just never seem to get to them | 21:07 |
| acoles | I marked the cross-compat test patch as WIP: not all tests pass with Swift s3api | 21:07 |
| timburke | i've been poking at them as something detached from the chain, and have it down from 8 failures to 5 -- we'll see what i can do for the rest | 21:08 |
| acoles | I don't plan to immediately address those failures (they're not data-path related discrepancies) | 21:08 |
| acoles | Instead, I'm trying to make progress towards an all-new swift s3api MPU implementation, but it's very early days | 21:09 |
| timburke | it'll be great :D | 21:09 |
| timburke | while i'm still on the topic of s3 compat, i still haven't done much with https://review.opendev.org/c/openstack/swift/+/836755 but want to dig in more over the next week | 21:11 |
| patch-bot | patch 836755 - swift - Add support of Sigv4-streaming - 8 patch sets | 21:11 |
| timburke | #topic py312 | 21:12 |
| timburke | acoles merged https://review.opendev.org/c/openstack/swift/+/904601 (thanks!) | 21:12 |
| patch-bot | patch 904601 - swift - tests: Fix float expectations for py312 (MERGED) - 5 patch sets | 21:12 |
| acoles | 👌 | 21:13 |
| mattoliver | Speaking of py312, when having 312 as your default tox pep8 fails as it pulls in a later version of something, pinning it doesn't fix it. | 21:14 |
| timburke | and i think we're lined up to continue pushing on https://review.opendev.org/c/openstack/swift/+/904652 (which may help with some other issues we've seen on py37 🤞) | 21:14 |
| patch-bot | patch 904652 - swift - Add ClosingIterator class; be more explicit about ... - 7 patch sets | 21:14 |
| mattoliver | importlib-metadata > 5 removed an api that flake8 uses. | 21:15 |
| timburke | mattoliver, oh boo -- i think i hadn't tried linting yet, just tests | 21:15 |
| mattoliver | I could set a basepython to <312 in tox and it fixes it. | 21:15 |
| mattoliver | But that isn't great. | 21:15 |
| mattoliver | So might need to check on the status of flake8 upstream and see if they're working on using the new importlib-metadata api | 21:16 |
| mattoliver | raise an issue or maybe even fix it myself 🤷 | 21:16 |
| timburke | i wonder if maybe it's down to the version pins hacking uses 🤔 | 21:18 |
| mattoliver | Because my default python is 312 (thanks to fedora) I might continue chasing this down because it bugs and effects me most. | 21:19 |
| timburke | i also remembered recently that there's 312 work outside of swift -- we want https://review.opendev.org/c/openstack/pyeclib/+/839643 to get pyeclib tests working | 21:19 |
| patch-bot | patch 839643 - pyeclib - Drop support for liberasurecode<1.4.0 - 10 patch sets | 21:19 |
| mattoliver | maybe, I went in an forced a pin to the right version, but then it complained about wanting 2 versions of the module.. so yeah either something else is pulling it in, or is packaged/coming with my py312 | 21:20 |
| mattoliver | oh yeah. I forget about those repos. | 21:20 |
| timburke | (because of the deprecation called out in the commit message; py312 made good on the promise to remove) | 21:20 |
| timburke | but it'll also be really nice to finally close #1639691! | 21:21 |
| patch-bot | https://bugs.launchpad.net/swift/+bug/1639691 - EC: Swift can return corrupted Data and be able to go data lost at isa_l_rs_vand policy with >=5 parities (Fix Released) | 21:21 |
| timburke | i think that's it for updates from last week | 21:22 |
| timburke | #topic open discussion | 21:23 |
| timburke | anything else we want to bring up here? | 21:23 |
| acoles | not from me | 21:23 |
| mattoliver | Thanks to acoles p 809969 has been reworked and much better | 21:24 |
| patch-bot | https://review.opendev.org/c/openstack/swift/+/809969 - swift - sharding: don't replace own_shard_range without an... - 15 patch sets | 21:24 |
| acoles | I wonder if we'll get any hint as to the mystery epoch-resetting bug? | 21:24 |
| mattoliver | yeah no idea. I mean the reason we added device name to db_ids was a hope to find the culprit. | 21:25 |
| mattoliver | Also now that the namespace and container auto get patches have landed. I'm reworking p 877584 | 21:25 |
| patch-bot | https://review.opendev.org/c/openstack/swift/+/877584 - swift - internal_client: Add iter_{shard_ranges,namespaces... - 12 patch sets | 21:25 |
| mattoliver | Still working it through, but there is now an iter_shardranges and iter_namespace that do to the backends. And now also have a iter_namespaces_from_cache which only attempt to pull from cache (if enabled). Return nothing on cache_miss | 21:26 |
| mattoliver | which will be useful for smartening up the object updater | 21:27 |
| mattoliver | Last night I had to make dinner so quickly pushed up what I had. Almost have a more polished one to push up now. maybe after the meeting | 21:28 |
| timburke | i like smarter daemons :-) | 21:28 |
| timburke | all right, i think we'll call it early then | 21:29 |
| mattoliver | oh and I still have p 905064 lingering, but thats probably more on me :) | 21:29 |
| patch-bot | https://review.opendev.org/c/openstack/swift/+/905064 - swift - wip: shard replication sync points - 5 patch sets | 21:29 |
| mattoliver | yup, that's all I got :) | 21:29 |
| timburke | thanks for coming, and thank you for working on swift! | 21:29 |
| timburke | #endmeeting | 21:29 |
| opendevmeet | Meeting ended Wed Feb 7 21:29:33 2024 UTC. Information about MeetBot at http://wiki.debian.org/MeetBot . (v 0.1.4) | 21:29 |
| opendevmeet | Minutes: https://meetings.opendev.org/meetings/swift/2024/swift.2024-02-07-21.00.html | 21:29 |
| opendevmeet | Minutes (text): https://meetings.opendev.org/meetings/swift/2024/swift.2024-02-07-21.00.txt | 21:29 |
| opendevmeet | Log: https://meetings.opendev.org/meetings/swift/2024/swift.2024-02-07-21.00.log.html | 21:29 |
| timburke | oh, i forgot to mention: while i was reworking p 839649 (and refreshing my memory of how multi-region replication over ssync works) i was reminded of how terrible it is to try to use REPLICATE via curl, so i pushed up a fresh version of p 758638 | 21:32 |
| patch-bot | https://review.opendev.org/c/openstack/swift/+/839649 - swift - replicator: Add sync_batches_per_revert option - 14 patch sets | 21:32 |
| patch-bot | https://review.opendev.org/c/openstack/swift/+/758638 - swift - Be willing to send hashes as JSON - 10 patch sets | 21:32 |
| mattoliver | kk | 21:33 |
| opendevreview | Tim Burke proposed openstack/swift master: sharding: don't replace own_shard_range without an epoch https://review.opendev.org/c/openstack/swift/+/809969 | 21:38 |
| opendevreview | Anish Kachinthaya proposed openstack/swift master: add container level grace period https://review.opendev.org/c/openstack/swift/+/907762 | 22:23 |
| opendevreview | Merged openstack/swift master: replicator: Rename update_deleted to revert https://review.opendev.org/c/openstack/swift/+/908252 | 22:57 |
| opendevreview | Tim Burke proposed openstack/swift master: lint: Up-rev hacking https://review.opendev.org/c/openstack/swift/+/908367 | 22:59 |
| opendevreview | Tim Burke proposed openstack/swift master: lint: Consistently use assertIsInstance https://review.opendev.org/c/openstack/swift/+/908368 | 22:59 |
| timburke | mattoliver, see what you think of the hacking up-rev; i think it should resolve your py312 issue ^^^ | 23:00 |
| timburke | boo! starting with hacking 5.0+, it requires py38+... | 23:21 |
| clarkb | you should be able to run hacking with newer python and not have that affect the versions of python you support at runtime though | 23:25 |
| mattoliver | Kk | 23:28 |
| mattoliver | Lol it never ends 😜 | 23:28 |
| opendevreview | Tim Burke proposed openstack/swift master: lint: Up-rev hacking https://review.opendev.org/c/openstack/swift/+/908367 | 23:48 |
| opendevreview | Tim Burke proposed openstack/swift master: lint: Consistently use assertIsInstance https://review.opendev.org/c/openstack/swift/+/908368 | 23:48 |
Generated by irclog2html.py 2.17.3 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!