Adjusting garbage collection for delete-heavy workloads

Delete-heavy workloads can outpace garbage collection when large numbers of short-lived objects are created and removed in a short time.

In some environments, garbage collection activity may not keep up with the rate at which objects are deleted, either temporarily or on an ongoing basis. This situation commonly occurs in delete-heavy workloads, where objects are stored briefly and then removed. In these cases, increasing the priority of garbage collection operations relative to other operations can help maintain system stability. For additional questions about Ceph Object Gateway garbage collection, contact IBM Support.

Prerequisites

  • A running IBM Storage Ceph cluster.

  • Root-level access to all nodes in the storage cluster.

Procedure

  1. Set the value of rgw_gc_max_concurrent_io to 20, and the value of rgw_gc_max_trim_chunk to 64:

    Example 

    [ceph: root@host01 /]# ceph config set client.rgw rgw_gc_max_concurrent_io 20
[ceph: root@host01 /]# ceph config set client.rgw rgw_gc_max_trim_chunk 64

  2. Restart the Ceph Object Gateway to allow the changed settings to take effect.

  3. Monitor the storage cluster during GC activity to verify that the increased values do not adversely affect performance.

Important: Never modify the value for the rgw_gc_max_objs option in a running cluster. You should only change this value before deploying the RGW nodes.