29 Oct
2024
29 Oct
'24
2:11 a.m.
Did you bulk delete a lot of volumes? Are you sure there aren't any snapshots involved? The math for the 'rbd du' output is plausible: NAME PROVISIONED USED <TOTAL> 14 TiB 8.3 TiB 3.94 Million objects of size 4M results in roughly 14 or 15 TB, so it doesn't seem like there are orphaned objects. Can you share 'ceph osd df', 'ceph -s' and 'ceph pg ls-by-pool volumes' in a text file? BTW, replica size 2 is a really bad choice, it has been discussed many times why it should only be considered in test clusters or if your data is not important. Zitat von lfsilva@binario.cloud:
Hello Eugen,
Yes, I ran the command but there is no output:
"root@hc-node01:~# rbd -p volumes trash ls root@hc-node01:~# "