warning: fast-diff map is invalid for vm

[abzsol]

Hi, if i do a rbd du i have this warning, each for every snapshot and the image itself.

warning: fast-diff map is invalid for vm-110-disk-3

i noticed that on a newer vms this not happen, so i think is something related to new versions of cecph. I started the cluster with version 5.2, 2 years ago, and then updated to 6 and now 6.2 (following instruction for pve5to6)

if a do a rbd-objectmap-rebuild the error disappear.

Is it the correct way? i can i "script" and do this in batch for all vms and relative snapshots?
i've tried with

rbd ls cephpool | rbd object-map rebuild

but it does not work

Thanks.

 

[wolfgang]

Hi,

I guess it has something to do with image features.
can you please send the output of

rbd info -p <poolname(default rbd)> vm-110-disk-3

 

[abzsol]

Hi,

I guess it has something to do with image features.
can you please send the output of

rbd info -p <poolname(default rbd)> vm-110-disk-3

rbd image 'vm-110-disk-3':
size 3.9 TiB in 1022362 objects
order 22 (4 MiB objects)
snapshot_count: 14
id: 25d9c06b8b4567
block_name_prefix: rbd_data.25d9c06b8b4567
format: 2
features: layering, exclusive-lock, object-map, fast-diff
op_features:
flags:
create_timestamp: Fri Feb 8 12:55:38 2019


i did the commands manually for image and all the snapshot. and from this problems started, i don't know if it's related:


https://forum.proxmox.com/threads/ceph-storage-all-pgs-snaptrim-every-night-slowing-down-vms.71573/

 

[wolfgang]

The problem is that your image have no "deep flatten" feature.
This is an immutable feature and can't be turned on.
A solution would be to copy the image but then you lose all snapshots of it.

 

[abzsol]

The problem is that your image have no "deep flatten" feature.
This is an immutable feature and can't be turned on.
A solution would be to copy the image but then you lose all snapshots of it.

ok. i tried creating a new disk and this new disk has deep flatten enabled. i'll move all files in this new image.

Could this the root cause of this? https://forum.proxmox.com/threads/ceph-storage-all-pgs-snaptrim-every-night-slowing-down-vms.71573/