Administrator Guide

Even though the internal snapshots are hidden from the user, they do consume snapshot space (and thus pool space) from the virtual
pool. If the volume is the base volume for a snapshot tree, the count of maximum snapshots in the snapshot tree may include the internal
snapshots for it even though they are not listed. Internal snapshots and internal volume groups count against system limits, but do not
display.
Creating a virtual pool for replication
When you create a virtual pool, specify that it has enough space for three times the anticipated size of the primary volume to account for
the primary volume plus the same amount of space for each of the two internal snapshots. This is the maximum amount of space that you
will need for replication. Also, for a pool on the primary system, allow additional space for other uses of the pool.
Setting up snapshot space management in the context of
replication
The snapshot space management feature, accessible only through the CLI, enables users to monitor and control the amount of space that
snapshots can consume in a pool. In addition to configuring a snapshot space limit, you can also specify a limit policy to enact when the
snapshot space reaches the configured limit. The policy will either notify you via the event log that the percentage has been reached (in
which case the system continues to take snapshots, using the general pool space), or notify you and trigger automatic deletion of
snapshots. If automatic deletion is triggered, snapshots are deleted according to their configured retention priority. For more information
on setting snapshot retention priority, see Maintaining replication snapshot history from the Replications topic.
When you create virtual volumes through the create volume and create volume-set CLI commands, you can set the retention
priority for snapshots of the volume. If automatic deletion of snapshots is enabled, the system uses the retention priority of snapshots to
determine which, if any, snapshots to delete. Snapshots are considered to be eligible for deletion if they have any retention priority other
than never-delete. Snapshots are configured to be eligible for deletion by priority and age. The oldest, lowest priority snapshots are
deleted first. Internal replication snapshots and snapshots that are mapped or are not leaves of a volume's snapshot tree are ineligible for
deletion. For more information on the create volume and create volume-set CLI commands, see the Dell EMC PowerVault ME4
Series Storage System CLI Guide.
If you are using the replication feature and snapshot space management, there are specific factors to consider when managing snapshot
space for the primary and secondary systems, especially when setting up the snapshot space and policies for the pool:
• Make sure that there is enough snapshot space to accommodate the maximum anticipated size of the two internal snapshots, which
cannot be deleted, and any other snapshots that you would like to retain.
• To adjust the snapshot space of the pool, increase the value of the limit parameter of the set snapshot-space CLI command. For
more information on the set snapshot-space CLI command, see the Dell EMC PowerVault ME4 Series Storage System CLI
Guide.
• You can later create more snapshot space by adding disks to the pool to increase its size.
If the internal snapshots are larger than anticipated and take up a lot of snapshot space, you can adjust the snapshot space thresholds or
increase the snapshot space to prevent unintentional automatic deletion of snapshots that you want to retain. To monitor the snapshot
space for virtual pools, use the show snapshot-space CLI command. To monitor the size of the internal snapshots, use the show
snapshots CLI command with its type parameter set to replication. For more information on the show snapshots CLI command,
see the Dell EMC PowerVault ME4 Series Storage System CLI Guide.
Replication and empty allocated pages
Deleting data from a volume can result in deallocation of pages on that volume. Pages deallocated before the initial replication will not be
copied to the secondary volume. Pages deallocated since the last replication cause a page consisting of zeroes to be written to the
secondary volume during replication. This can result in a difference in the number of allocated pages between the primary and secondary
volumes. A virtual storage background task automatically reclaims pages consisting of all zeroes, eventually freeing up the secondary
volume snapshot space that these reclaimed pages consumed. Freeing up this space is not immediate and happens over a period of time.
Disaster recovery
The replication feature supports manual disaster recovery only. It is not integrated with third-party disaster recovery software. Since
replication sets of virtual volumes cannot reverse the direction of the replication, carefully consider how the replicated data will be
accessed at the secondary backup site when a disaster occurs.
Working in the Replications topic
115