Confusing CEPH PG warrning

[beriapl]

Hi,

I have setup with 3 physical hosts, each of them got 3 disks as OSD in CEPH cluster.

Using calculation:

Total PGs = (Total_number_of_OSD * 100) / max_replication_count

Where my max_replication_count = 3

I've set PG 128 for my POOL_CEPH

I've two pools:

rados lspools
device_health_metrics
POOL_CEPH

Max PG per OSD is set to 250.

ceph --admin-daemon /var/run/ceph/ceph-mon.pve11.asok config get  mon_max_pg_per_osd
{
    "mon_max_pg_per_osd": "250"
}

Pool device_health_metrics which is created by itself (during CEPH configuration) have 1 PG, while pool that I've created POOL_CPEH got 128 PG.

I'm getting warning 1 pools have too many placement groups :

ceph status
  cluster:
    id:     bf79845c-f78b-4b28-8bf9-85fb8d320a38
    health: HEALTH_WARN
            1 pools have too many placement groups

  services:
    mon: 3 daemons, quorum pve11,pve12,pve13 (age 12d)
    mgr: pve12(active, since 12d), standbys: pve11, pve13
    osd: 9 osds: 9 up (since 12d), 9 in (since 12d)

  data:
    pools:   2 pools, 129 pgs
    objects: 101.33k objects, 394 GiB
    usage:   1.2 TiB used, 6.7 TiB / 7.9 TiB avail
    pgs:     129 active+clean

What I'm missing to get rid of that warning? It doesn't exist with setup 64 PG for my pool.

 

[fabian]

https://docs.ceph.com/en/latest/rados/operations/placement-groups/