CEPH "22 OSD(s) have broken BlueStore compression.

[AngryAdm]

A nice little yellow warning appeared on my Ceph pool after having enabled and subsequently disabled lz4 compression.

What does this mean? Pool runs "fine" but, how do I get rid of this error?

22 OSD(s) have broken BlueStore compression osd.0 unable to load:none
osd.1 unable to load:none
osd.2 unable to load:none
osd.3 unable to load:none
osd.4 unable to load:none
osd.5 unable to load:none
osd.6 unable to load:none
osd.7 unable to load:none
osd.8 unable to load:none
osd.9 unable to load:none
osd.10 unable to load:none
osd.11 unable to load:none
osd.12 unable to load:none
osd.13 unable to load:none
osd.14 unable to load:none
osd.15 unable to load:none
osd.16 unable to load:none
osd.17 unable to load:none
osd.18 unable to load:none
osd.19 unable to load:none
osd.20 unable to load:none
osd.21 unable to load:none

 

[skliarie]

For those who came here by google

A bit cumbersome, but doable: unset compression_mode for each pool and restart each ceph-osd daemon:

#First unset the compression_mode in all pools
for pool in $(ceph osd lspools | awk '{print $2}');
{
ceph osd pool set $t compression_mode none
}

# then restart osd daemons - needs to be done on each machine
root@prox2:# systemctl list-units | grep ceph-osd@
  ceph-osd@10.service                                                                         loaded active     running   Ceph object storage daemon osd.10
  ceph-osd@11.service                                                                         loaded active     running   Ceph object storage daemon osd.11
  ceph-osd@12.service                                                                         loaded active     running   Ceph object storage daemon osd.12
  ceph-osd@13.service                                                                         loaded active     running   Ceph object storage daemon osd.13
  ceph-osd@14.service                                                                         loaded active     running   Ceph object storage daemon osd.14
  ceph-osd@15.service                                                                         loaded active     running   Ceph object storage daemon osd.15
  ceph-osd@5.service                                                                          loaded active     running   Ceph object storage daemon osd.5
...

root@prox2:# s=0; for t in 11 12 13 14 15 5; { sleep $s; systemctl restart ceph-osd@${t}.service; s=60; }