I don't think it's a memory leak anymore. But I created a tracker:
https://tracker.ceph.com/issues/69885
Zitat von Eugen Block :
To me it looks like a memory leak which wasn't present in 16.2.11
(the previous Ceph version on this cluster). The usage hasn't
changed, so it must be Ceph. I've
To me it looks like a memory leak which wasn't present in 16.2.11 (the
previous Ceph version on this cluster). The usage hasn't changed, so
it must be Ceph. I've been watching podman stats for a bit, the rgw
process uses more and more memory, until it caps at around 5 or 6 GB,
then it respa
