Hi Cephers,

These are the topics that we just covered in today's meeting:

Issues recording our meetings in Jitsi (Mike Perez)

David Orman suggested using a self-hosted Jitsi instance: https://jitsi.github.io/handbook/docs/devops-guide/. Tested on a single container, with 4 cores for 4-5 attendees a call.
Help needed from Dan Mick & Adam Kraitman.
Ilya: this might scale for smaller meetings (dailies), but not for the larger ones (CDM, User-Devs, etc.).

Reef release status review

Josh: CentOS 9 blocked by Dashboard/Cephadm Python deps missing in CentOS 9. Casey: It's blocking teuthology testing.

Already discussed in devs mailing list a couple of months ago, for Quincy + Centos 9 (the missing packages would be the same).
Ken Dreyer preferred to keep the legacy approach (distro packages) instead of embedding Python deps.
Casey, Matt & Ernesto to resume that discussion with Ken Dreyer.

Other issues:

Radek:

Performance issue with RocksDB config.
Mismatching client-server features during upgrade. msgr encoder-decoder issues, introduced feature bit for Squid (https://github.com/ceph/ceph/commit/1049d3e5eff0b7fa4fc9e5853494cb21c10b290a). Performance concerns. To be further discussed at Mark's Perf meeting (incl. Yuval)

Paul Cuzner perf testing Reef: higher CPU usage in Reef vs Quincy (more time spent in RocksDB get calls)

Target GA remains this June. Missing CentOS 9 packages becoming a blocker issue for the release.

Perf regression in Pacific vs Nautilus (David Orman)

https://tracker.ceph.com/issues/58530
Mark: Missing change in upstream Rocksdb project.
David Orman: Is this degradation still happening in newer Rocksdb versions (reef)? Mark: No reason to think otherwise.

Kind Regards,

Ernesto