.Upgrading to the latest version no longer causes cluster data movement
Previously, upgrading a {product} cluster to the latest version when CRUSH device classes were enabled, the `crushtool` utility rebalanced data in the cluster because of changes in the CRUSH map. This data movement should not have occurred. With this update, a reclassify functionality is available to help transition from older CRUSH maps that maintains parallel hierarchies for OSDs of different types to a modern CRUSHmap that makes use of the device class feature without triggering data movement.
Description of problem:
Currently when a legacy cluster is upgraded to Luminous (RHCS 3.y) and CRUSH device classes are enabled CRUSH will rebalance data in the cluster due to a change in the CRUSH map.
Reclassifying devices should not incur a data movement as we are only applying a label.
Upstream PR:
https://github.com/ceph/ceph/pull/24502
Version-Release number of selected component (if applicable):
3.y
Created attachment 1496814[details]
updated crush map
Updated CRUSH map that uses hdd-sas and hdd-sata classes. All mappings should be identical to before (no PGs should move).
Since the problem described in this bug report should be
resolved in a recent advisory, it has been closed with a
resolution of ERRATA.
For information on the advisory, and where to find the updated
files, follow the link below.
If the solution does not work for you, open a new bug report.
https://access.redhat.com/errata/RHSA-2019:0911