On 2020-08-28 12:57, Martin Palma wrote: > Thanks for the clarification. And by raising the PG should we also set > "no scrub" and "no deep scrub" during data movement? What is here the > recommandation? Yes, I would do that. We have seen that after creating more PGs those PGs will get (deep-)scrubbed as they have never been (deep-)scrubbed before. Only enable (deep-)scubbing after all data has been moved and cluster is HEALTH_OK. Because it will take a long time before they are all (deep-)scrubbed. Make sure you have a "scrub window" that is convenient (low traffic). Quite a few PGs might get scrubbed on different OSDs at the same time. Reminds me for the need for a "scrub" module in the manager to handle this stuff properly ... as soon as I have some time I'll come up with a proposal. Gr. Stefan _______________________________________________ ceph-users mailing list -- ceph-users@xxxxxxx To unsubscribe send an email to ceph-users-leave@xxxxxxx
 |