Data integrity and the use of consistency groups

Disk based replication technologies provide a crash consistent copy of the replicated data. This means that the data that is not committed to the second site when a failure occurs may not be saved. However, a volume can be recovered to a recent point in time when all of the data was consistent.

A group of volumes that share a common recovery point is commonly called a consistency group. The storage controller ensures that after a failure, all of the volumes within a consistency group are recovered to the same point in time.

When using a storage-based replication with IBM Storage Scale, it is important to ensure that all the NSD’s in a file system are contained within the same consistency group. This way the metadata NSD’s are always in sync with the data NSD’s after a failure.