Den ons 18 nov. 2020 kl 04:59 skrev Szabo, Istvan (Agoda) < Istvan.Szabo@xxxxxxxxx>: > I have this error: > I have 36 osd and get this: > Error ERANGE: pg_num 4096 size 6 would mean 25011 total pgs, which > exceeds max 10500 (mon_max_pg_per_osd 250 * num_in_osds 42) > I have 4:2 data EC pool, and the others are replicated. > > pool 12 'sin.rgw.buckets.data' erasure profile data-ec size 6 min_size 5 > crush_rule 3 object_hash rjenkins pg_num 32 pgp_num 32 autoscale_mode warn > last_change 604 flags hashpspool,ec_overwrites stripe_width 16384 > application rgw > Don't forget EC 4+2 means every PG in the pool becomes 4+2 PGs out on the OSDs, so 4096*6 is 25k, and if you aim for something like the recommended 100-200 PGs per OSD, you should sum the PGs (taking replication factor and K+M size into account!) and try to end up at 3600-7200 PGs as the sum for ALL pools. Setting 4096 on a size 6 pool would in itself go way past this alone, regardless of the other pools. -- May the most significant bit of your life be positive. _______________________________________________ ceph-users mailing list -- ceph-users@xxxxxxx To unsubscribe send an email to ceph-users-leave@xxxxxxx
 |