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
_______________________________________________
ceph-users mailing list -- [email protected]
To unsubscribe send an email to [email protected]
