[SOLVED] Ceph pool is not empty

[mapachito1]

Hi,

I want to remove a pool (ceph-xxx pool), then I move vms to anothers pools (backup and restore in new pool). The problem is that the pool is not empty...

ceph df
--- RAW STORAGE ---
CLASS     SIZE    AVAIL     USED  RAW USED  %RAW USED
hdd    3.9 TiB  3.2 TiB  697 GiB   697 GiB      17.41
ssd    5.6 TiB  4.1 TiB  1.4 TiB   1.4 TiB      25.79
TOTAL  9.5 TiB  7.4 TiB  2.1 TiB   2.1 TiB      22.33

--- POOLS ---
POOL      ID  PGS   STORED  OBJECTS     USED  %USED  MAX AVAIL
ceph-xxx   2   32  122 GiB   31.52k  367 GiB   6.33    1.8 TiB
.mgr       3    1   66 MiB       14  199 MiB      0    1.8 TiB
ceph-ssd   6  128  355 GiB   91.67k  1.0 TiB  25.07    1.0 TiB
ceph-hdd   7  128  238 GiB   62.93k  713 GiB  19.28    995 GiB

I have verified that I don't have any vm in ceph-xxx,

How can I know why it is not empty? If I do

rados -p ceph-xxx ls

I read a lot rbd_data* objects

 

[gurubert]

What does "rbd -p ceph-xxx ls -l" output?

 

[mapachito1]

What does "rbd -p ceph-xxx ls -l" output?

rbd -p ceph-xxx ls -l
NAME           SIZE    PARENT  FMT  PROT  LOCK
vm-244-disk-0  32 GiB            2
vm-301-disk-0  60 GiB            2
vm-302-disk-0  64 GiB            2

mmmm, I don't have those vms in ceph-xxx. For example 301....

cat /etc/pve/nodes/node2/qemu-server/301.conf
boot: order=sata0
cores: 1
hotplug: disk
ide2: none,media=cdrom
...
sata0: ceph-ssd:vm-301-disk-0,size=60G
scsihw: virtio-scsi-single
...

Previously this vm has ceph-xxx but I changed it. If I launch

grep -r 'ceph-xxx' /etc/pve/nodes

I don't have any response. Maybe trash? How can I purge it?

 

[gurubert]

Then you did a storage migration and the source image was not deleted.
If you are sure that these images are not needed any more you can just delete the pool.

 

[mapachito1]

Then you did a storage migration and the source image was not deleted.
If you are sure that these images are not needed any more you can just delete the pool.

Yes, I have remove it and apparently I don't have any error.
Thks