Hi,
assuming you deployed with cephadm since you're mentioning Octopus
there's a brief section in [1]. The basis for the OSD deployment is
the drive_group configuration. If nothing has changed in your setup
and you replace an OSD cephadm will detect the available disk and
match it with the drive_group config. If there's enough space on the
SSD too, it will redeploy the OSD.
The same goes for your second case: you'll need to remove all OSDs
from that host, zap the devices, replace the SSD and then cephadm will
deploy the entire host. That's the simple case. If redeploying all
OSDs on that host is not an option you'll probably have to pause
orchestrator in order to migrate devices yourself to prevent to much
data movement.
Regards,
Eugen
[1] https://docs.ceph.com/en/latest/mgr/orchestrator/#replace-an-osd
Zitat von Tony Liu <tonyliu0592@xxxxxxxxxxx>:
Hi,
I did some search about replacing osd, and found some different
steps, probably for different release?
Is there recommended process to replace an osd with Octopus?
Two cases here:
1) replace HDD whose WAL and DB are on a SSD.
1-1) failed disk is replaced by the same model.
1-2) working disk is replaced by bigger one.
2) replace the SSD holding WAL and DB for multiple HDDs.
Thanks!
Tony
_______________________________________________
ceph-users mailing list -- ceph-users@xxxxxxx
To unsubscribe send an email to ceph-users-leave@xxxxxxx
_______________________________________________
ceph-users mailing list -- ceph-users@xxxxxxx
To unsubscribe send an email to ceph-users-leave@xxxxxxx
