Hi, I assume you're still on a "low" pacific release? This was fixed by PR [1][2] and the warning is supressed when autoscaler is on, it was merged into Pacific 16.2.8 [3]. I can't answer why autoscaler doesn't increase the pg_num, but yes, you can increase it by yourself. The pool for cephfs_metadata should be on fast storage and doesn't have huge amounts of data so it should be relatively quick. What's your 'ceph osd df' output? Regards, Eugen [1] https://tracker.ceph.com/issues/53644 [2] https://github.com/ceph/ceph/pull/45152 [3] https://docs.ceph.com/en/latest/releases/pacific/#v16-2-8-pacific Zitat von Edouard FAZENDA <e.fazenda@csti.ch>:

Hello Ceph community,

I have since this morning a warning about MANY_OBJECT_PER_PG on 1 pool which is cephfs_metadata

# ceph health detail

HEALTH_WARN 1 pools have many more objects per pg than average

[WRN] MANY_OBJECTS_PER_PG: 1 pools have many more objects per pg than average

pool cephfs_metadata objects per pg (154151) is more than 10.0215 times cluster average (15382)

I have the autoscaling on on all the pool :

# ceph osd pool autoscale-status

POOL SIZE TARGET SIZE RATE RAW CAPACITY RATIO TARGET RATIO EFFECTIVE RATIO BIAS PG_NUM NEW PG_NUM AUTOSCALE

device_health_metrics 9523k 3.0 26827G 0.0000 1.0 1 on

cephfs_data 5389G 2.0 26827G 0.4018 1.0 512 on

cephfs_metadata 19365M 2.0 26827G 0.0014 4.0 16 on

.rgw.root 1323 3.0 26827G 0.0000 1.0 32 on

default.rgw.log 23552 3.0 26827G 0.0000 1.0 32 on

default.rgw.control 0 3.0 26827G 0.0000 1.0 32 on

default.rgw.meta 11911 3.0 26827G 0.0000 4.0 8 on

default.rgw.buckets.index 0 3.0 26827G 0.0000 4.0 8 on

default.rgw.buckets.data 497.0G 3.0 26827G 0.0556 1.0 32 on

kubernetes 177.2G 2.0 26827G 0.0132 1.0 32 on

default.rgw.buckets.non-ec 432 3.0 26827G 0.0000 1.0 32 on

Actually the pg_num is 16 for the cephfs_metdata pool , but it does not define NEW_PG_NUM

Here the replicated size of all my pool

# ceph osd dump | grep 'replicated size'

pool 1 'device_health_metrics' replicated size 3 min_size 2 crush_rule 0 object_hash rjenkins pg_num 1 pgp_num 1 autoscale_mode on last_change 189372 flags hashpspool stripe_width 0 pg_num_min 1 application mgr_devicehealth

pool 10 'cephfs_data' replicated size 2 min_size 1 crush_rule 1 object_hash rjenkins pg_num 512 pgp_num 512 autoscale_mode on last_change 189346 lfor 0/0/183690 flags hashpspool,selfmanaged_snaps stripe_width 0 application cephfs

pool 11 'cephfs_metadata' replicated size 2 min_size 1 crush_rule 1 object_hash rjenkins pg_num 16 pgp_num 16 autoscale_mode on last_change 187861 lfor 0/187861/187859 flags hashpspool stripe_width 0 pg_autoscale_bias 4 pg_num_min 16 recovery_priority 5 application cephfs

pool 18 '.rgw.root' replicated size 3 min_size 2 crush_rule 0 object_hash rjenkins pg_num 32 pgp_num 32 autoscale_mode on last_change 5265 flags hashpspool stripe_width 0 application rgw

pool 19 'default.rgw.log' replicated size 3 min_size 2 crush_rule 0 object_hash rjenkins pg_num 32 pgp_num 32 autoscale_mode on last_change 5267 flags hashpspool stripe_width 0 application rgw

pool 20 'default.rgw.control' replicated size 3 min_size 2 crush_rule 0 object_hash rjenkins pg_num 32 pgp_num 32 autoscale_mode on last_change 5269 flags hashpspool stripe_width 0 application rgw

pool 21 'default.rgw.meta' replicated size 3 min_size 2 crush_rule 0 object_hash rjenkins pg_num 8 pgp_num 8 autoscale_mode on last_change 5398 lfor 0/5398/5396 flags hashpspool stripe_width 0 pg_autoscale_bias 4 pg_num_min 8 application rgw

pool 22 'default.rgw.buckets.index' replicated size 3 min_size 2 crush_rule 0 object_hash rjenkins pg_num 8 pgp_num 8 autoscale_mode on last_change 7491 lfor 0/7491/7489 flags hashpspool stripe_width 0 pg_autoscale_bias 4 pg_num_min 8 application rgw

pool 23 'default.rgw.buckets.data' replicated size 3 min_size 2 crush_rule 0 object_hash rjenkins pg_num 32 pgp_num 32 autoscale_mode on last_change 7500 flags hashpspool stripe_width 0 application rgw

pool 24 'kubernetes' replicated size 2 min_size 1 crush_rule 1 object_hash rjenkins pg_num 32 pgp_num 32 autoscale_mode on last_change 189363 lfor 0/0/7560 flags hashpspool,selfmanaged_snaps stripe_width 0 application rbd

pool 25 'default.rgw.buckets.non-ec' replicated size 3 min_size 2 crush_rule 0 object_hash rjenkins pg_num 32 pgp_num 32 autoscale_mode on last_change 23983 flags hashpspool stripe_width 0 application rgw

Why the autoscaler is not acting to increase the pg_num of the pool in warning ?

As the pgcalc is not available on ceph now , do you thing it is a good idea to increase manually the pg_num of the cephfs_metadata , but which value should I set ?

I have 18 OSD

Thanks for the help

Best Regards,

Edouard FAZENDA

Technical Support

Chemin du Curé-Desclouds 2, CH-1226 THONEX +41 (0)22 869 04 40

<https://www.csti.ch/> www.csti.ch