meta data for this page
  •  

Differences

This shows you the differences between two versions of the page.

Link to this comparison view

Both sides previous revisionPrevious revision
vm:proxmox:ceph:pg [2023/07/03 20:37] – niziakvm:proxmox:ceph:pg [2023/07/11 12:06] (current) – niziak
Line 8: Line 8:
   * Max PG per OSD is 300   * Max PG per OSD is 300
   * Result must be rounded up to the nearest power of 2.   * Result must be rounded up to the nearest power of 2.
 +
 +===== reweight =====
 +
 +Adjust reweight according to current OSD utilisation - to prevent filling one OSD
 +<code bash>ceph osd reweight-by-utilization</code>
 +
 +In CRUSH hierarchies with a smaller number of OSDs, it’s possible for some OSDs 
 +to get more PGs than other OSDs, resulting in a higher load. You can reweight 
 +OSDs by PG distribution to address this situation by executing the following:
 +<code bash>ceph osd reweight-by-pg 101</code>
 +
 +NOTE: 101 is a ''oload'' parameter, which means ''overload percentage for consideration 101%''. For example, an argument of 200 would reweight OSDs which are twice as utilized as the average OSD.
 +
 +
 +===== Balancer =====
 +
 +When the cluster is healthy, the balancer will incrementally move a small fraction of unbalanced PGs in order to improve distribution. This fraction will not exceed a certain threshold that defaults to 5%.
 +Set it to 1%:
 +<code bash>ceph config set mgr target_max_misplaced_ratio .01</code>
 +