| opendevreview | Matthew Oliver proposed openstack/swift master: container-server: return objects of a given policy https://review.opendev.org/c/openstack/swift/+/803423 | 00:12 |
|---|---|---|
| opendevreview | Matthew Oliver proposed openstack/swift master: ring: Keep track of last primary nodes from last rebalance https://review.opendev.org/c/openstack/swift/+/790550 | 01:40 |
| opendevreview | Matthew Oliver proposed openstack/swift master: WIP: Reconstructor: Use past node and abort to handoff https://review.opendev.org/c/openstack/swift/+/792075 | 01:40 |
| opendevreview | Matthew Oliver proposed openstack/swift master: reconstructor: add skip_partition_count option https://review.opendev.org/c/openstack/swift/+/793888 | 01:40 |
| opendevreview | Matthew Oliver proposed openstack/swift master: ring: store actual ring replica count to better deserialize https://review.opendev.org/c/openstack/swift/+/803665 | 01:40 |
| mattoliver | timburke: split the ring change into 2. The first just fixes the deserialize over read issue which will allow us to start storing extra data in a ring file beyond the replica2part2dev_id structure (which is needed for storing the last primary array). | 01:43 |
| mattoliver | But having it seperate means if we merge it first and backport it, then any old swifts that at least run that patch would be able to open rings created by newer code, so helps with upgrade and potential downgrades. | 01:44 |
| timburke_ | i mean, there's old swifts and there's old swifts. no amount of backporting is going to make it so a node on 2.8.0 (say) can read this new format we introduce with 2.28.0+ -- i think there's probably no avoiding the need for a version bump on the format that'll prevent the old code from loading the new format | 02:06 |
| timburke_ | might be able to at least continue writing the old way if there's no fractional component, though | 02:07 |
| mattoliver | yeah, good point. | 02:42 |
| mattoliver | did some very basic initial benchmarking for the req tracking patch chain: https://etherpad.opendev.org/p/swift-req-tracking-getput-results | 07:25 |
| *** diablo_rojo is now known as Guest3447 | 09:16 | |
| opendevreview | Alistair Coles proposed openstack/swift master: container-server: return objects of a given policy https://review.opendev.org/c/openstack/swift/+/803423 | 10:48 |
| opendevreview | Alistair Coles proposed openstack/swift master: Preserve X-Backend-Storage-Policy-Index when recursively listing shards https://review.opendev.org/c/openstack/swift/+/803729 | 10:48 |
| timburke_ | mattoliver, nice! seems like there's little difference between "Master" and "Trace -- not tracing any request" which is exactly what we'd hope for | 15:47 |
| opendevreview | Pete Zaitcev proposed openstack/swift master: Make the dark data watcher work with sharded containers https://review.opendev.org/c/openstack/swift/+/787656 | 22:25 |
Generated by irclog2html.py 2.17.2 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!