Enabling non-blocking bucket resharding

Enable bucket index resharding in both single-site and multi-site deployments to improve RGW client performance.

Before you begin

Before you begin, make sure that you have the following prerequisites in place:
  • IBM Storage Ceph deployment uses the version that supports non-blocking resharding.
  • All RGW, rgw-admin, and OSD nodes in the cluster are upgraded to the same supported version.
  • Any ongoing bucket resharding operations are completed before performing the upgrade.

About this task

By default, bucket resharding temporarily blocks write and delete operations, especially on large buckets. Enabling non-blocking resharding allows write and delete operations to continue with minimal disruption while resharding happens in the background.
Note: Dynamic resharding is enabled by default in this release. Verify that it remains enabled.

Procedure

  1. Upgrade the IBM Storage Ceph cluster to a version that supports non-blocking resharding.
    Note: Do not perform a partial upgrade. All relevant components must be upgraded. Mismatched versions between components may cause resharding to fail or revert to blocking mode.
  2. Verify that all RGW, rgw-admin, and OSD nodes are running the same version.

Results

After the upgrade, bucket resharding occurs in the background without significantly disrupting write or delete operations. This improves performance during resharding events.