Resilient and non-resilient data pools

Resilient data pools enable data to replicate or encode its data. A non-resilient data pool does not enable to replicate or encode its data. Non-resilient pools are also known as replical.

A non-resilient pool is not protected from data loss. A small cluster with non-resilient data pools does its own replication and does not need replication from IBM Storage Ceph as it does its own replication.

Cluster utilization with data pools

With smaller configurations, the cluster utilization reduces because of the loss of resiliency. The recovery is limited by host utilization and might potentially impact production input/output operations per second (IOPS). When there is a single node of failure, you can add a third node to limit the host utilization and for IBM Storage Ceph to recover to full replication.