HI Nigel, In Nautilus you can decrease PG , but it take weeks , for example for us to go from 4096 to 2048 took more than 2 weeks. First at all pg-autoscaling is activable by pool. And you’re going to get a lot of warning , but it works. Normally is recommended upgrade a cluster with HEALTH_OK state. Also is recommended to use the unmap method the get the perfect distribution at balancer module, but it don’t work with misplaced/degraded error states. From my poin of view I will try go Healthy , them upgrade. Remember that you MUST repair all your SSD pre-nautilus due statistics scheme changed. Regards Manuel De: ceph-users <ceph-users-bounces@xxxxxxxxxxxxxx> En nombre de Nigel Williams Due to a gross miscalculation several years ago I set way too many PGs for our original Hammer cluster. We've lived with it ever since, but now we are on Luminous, changes result in stuck-requests and balancing problems. The cluster currently has 12% misplaced, and is grinding to re-balance but is unusable to clients (even with osd_max_pg_per_osd_hard_ratio set to 32, and mon_max_pg_per_osd set to 1000). Can I safely press on upgrading to Nautilus in this state so I can enable the pg-autoscaling to finally fix the problem? thanks. |
_______________________________________________ ceph-users mailing list ceph-users@xxxxxxxxxxxxxx http://lists.ceph.com/listinfo.cgi/ceph-users-ceph.com
