Dell PowerFlex 4.6.x Technical Overview

Replication consistency group

The replication consistency group is a logical container for volumes whose application data must be replicated consistently to each other.

For example, if you replicate a database and its transaction log, you need the remote copy of each of the database volumes to reflect an image from the same point in time. You can achieve this by placing all the database volumes in one replication consistency group.

Replication is always defined in the scope of a protection domain. All the objects that participate in the replication are contained in the protection domain, including the volumes in a replication consistency group. The journal capacity from the storage pool in a protection domain is shared among all the replication consistency groups in the protection domain. The SDRs in the protection domain manage I/Os directed to replicated volumes within the protection domain.

The following replication attributes are among those that are defined in the replication consistency group:

• Recovery point objective — The maximum data loss (in units of time) that you are willing to lose. When you set the recovery point objective, this is the goal of replication.
• Direction — Local to remote or remote to local. When there is no replication, for example, following failover or switchover, the direction attribute indicates the direction of the last replication.
• Abstract state — State of replication, for which the options are as follows:
  • OK — Replication is healthy, and the recovery point objective target is met.
  • Recovery point objective violation — Replication is healthy, but the recovery point objective target is not met.
  • Error — Replication is not healthy. The Error state attribute provides additional information about the cause for the error.
  • Stopped by user — Replication is not carried out due to user action, such as pause or terminate.
• Error state — Describes the error that is preventing data replication.
• Pause mode — When the replication consistency group is paused, all application I/Os are stored in the source journal. When replication of the replication consistency group is resumed, the source SDR sends the journal contents to the target SDR to be applied to the target volumes. You may want to pause an replication consistency group to handle a network issue between the peer systems or when fixing a hardware issue.
• Activity state — When the replication consistency group is active, replication is enabled. When the replication consistency group is inactive, replication is stopped, but replication consistency group configuration is maintained. The replication consistency group can be automatically inactivated by the system (due to insufficient journal capacity), or can be terminated manually by the user. When activating an inactive replication consistency group, the replication consistency group starts replication from the beginning with a full initialization.
• Failover type — The failover operation requested by the user. For more information, see Accessing target volumes.
• Failover state — The states are transient, except for None (no failover) or Done (the failover operation has completed).