Hi, I have a 4:2 erasure coded data pool which based on host and have 6 data nodes. I've increased the pg from 32 to 128 but the recovery operation is less than 10MB/s. The server has bonded 25GB nic cards. In octopus I don't have to initiate the pgp_increase because it automatically adjust up, but at the moment we are just 45 pg out of the 128. Now started to have pg not deep scrubbed and I can't deep scrub them I guess because of data rebalance. What's the best thing to speed it up or scrub the pg? Of course these are set to low but even if it is low normally it should go up to GBs at least. osd_max_backfills = 1 osd_recovery_max_active = 1 osd_recovery_op_priority = 1 Thanks _______________________________________________ ceph-users mailing list -- ceph-users@xxxxxxx To unsubscribe send an email to ceph-users-leave@xxxxxxx
 |