| opendevreview | Alistair Coles proposed openstack/swift master: WIP proxy: Add systags abstraction for etag override https://review.opendev.org/c/openstack/swift/+/943034 | 13:36 |
|---|---|---|
| opendevreview | Alistair Coles proposed openstack/swift master: proxy: calculate md5 hash for override-etag when required https://review.opendev.org/c/openstack/swift/+/944040 | 13:36 |
| opendevreview | Shreeya Deshpande proposed openstack/swift master: test for cooperative token memcache https://review.opendev.org/c/openstack/swift/+/943055 | 14:49 |
| opendevreview | Shreeya Deshpande proposed openstack/swift master: Add labeled metrics to s3api https://review.opendev.org/c/openstack/swift/+/939481 | 15:38 |
| opendevreview | Merged openstack/swift master: Make FakeSwiftCall.DUMMY_VALUE a bespoke singleton https://review.opendev.org/c/openstack/swift/+/943684 | 15:53 |
| opendevreview | Tim Burke proposed openstack/swift master: Add magic rules for swift rings https://review.opendev.org/c/openstack/swift/+/943993 | 16:00 |
| DHE | if you're making a ring v2 format, I'd like to suggest a feature: recommended first hand-offs. get_more_nodes() would return these in the front of the new nodes list. during a rebalance, any nodes removed from a partition would be put into this "recommended handoffs" list. even if it's just 1 node per host. after min-part-hours passes a rebalance would also just remove that list item.. | 16:25 |
| clayg | @DHE that's a great idea, I think @mattoliver already wrote that code somewhere... and old_primaries table? I think you have to plumb a `get_more_nodes(for_read=True)` kind of thing.. | 17:34 |
| DHE | hmm.... as an outsider, I have mixed opinions on for_read=True.... I feel like it could cause an old version of data, including possibly a deleted state, to persist and delay availability... | 17:38 |
| opendevreview | Alistair Coles proposed openstack/swift master: Add support of Sigv4-streaming https://review.opendev.org/c/openstack/swift/+/836755 | 17:41 |
| opendevreview | Alistair Coles proposed openstack/swift master: utils: Add crc32c and crc64nvme functions https://review.opendev.org/c/openstack/swift/+/909800 | 17:41 |
| opendevreview | Alistair Coles proposed openstack/swift master: s3api: Add support for additional checksums https://review.opendev.org/c/openstack/swift/+/909801 | 17:41 |
| opendevreview | Tim Burke proposed openstack/swift master: s3api: Validate additional checksums on upload https://review.opendev.org/c/openstack/swift/+/944073 | 21:55 |
| opendevreview | Tim Burke proposed openstack/swift master: s3api: Validate additional checksums on upload https://review.opendev.org/c/openstack/swift/+/944073 | 22:05 |
| opendevreview | Shreeya Deshpande proposed openstack/swift master: test for cooperative token memcache https://review.opendev.org/c/openstack/swift/+/943055 | 22:10 |
| timburke_ | i was realizing that we probably want to go to old-primaries first for more than just reads -- POSTs (and maybe DELETEs? but there's override_statuses...) in particular would also benefit. maybe instead of for_read=True it should be idempotent_write=False or something? | 22:14 |
Generated by irclog2html.py 2.17.3 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!