Wednesday, 14 June 2017

Basic CEPH info.

Below list some state of CEPH heath and required action.

PG- Placement groups

active + clean = Normal state
        correct numbers of PG availbale
degrad =  fewer replicas available.
        No action needed, wait for some time. Basically happens when osd went down or removed.
        Can Execute recovery
recovering - catching up from previous degraded state.
backfill - is doing recovering.
        starts from empty replica.
        Fills new data on new disk.
incomplete = fewer replicas availbale.
        may need another OSD.
        may crush map change.
        change min_size for pool.
inconsistent = scrub or deep-scrub detects error.
        ceph pg repair required.
down - data missing. PG unavailable.
        declare the osd lost.

#ceph -s - shows the health status of CEPH cluster

# ceph heath - also show the health of cluster in shortly

#ceph osd tree -  gives you the status of OSDs

How to find out the disk is SSD or HDD ?
[root@osd02 ~]# cat /sys/block/sda/queue/rotational
if OP = 1 then it is HDD or if 0 then it is SSD

No comments:

Prasanta Kumar Behera. Mob- 8147470101