Release Notes

Linux setup and configuration
28 Dell EMC PowerVault ME4 Series and Linux | 3924-BP-L
3.6 Using snapshots for data protection
ME4 Series storage can create multiple point-in-time snapshots of virtual volumes. This section discusses
snapshots and some use cases.
Consider the following when deciding to use snapshots to protect data on virtual volumes.
Do not enable snapshots for all volumes. Carefully consider what data requires this type of protection
and plan accordingly. The maximum number of mappable snapshots per system is 1,024. When the
limit is reached, no snapshots can be created until existing snapshots are deleted.
If replication is enabled between two ME4 Series systems, the replication process creates internal
snapshots which are not visible to the administrator. These snapshots do not count against the per-
volume limit but do count toward the system limit.
Snapshots can be created manually, or created by a pre-defined schedule on a per-volume basis.
It is not possible to define a common snapshot schedule across multiple volumes. Therefore, if a
consistent snapshot is required for a group of volumes, manual coordination must be done from
applications and through the use of scripting to synchronize and automate the multi-step process.
Snapshots are treated like any other volumes. They can be mapped to hosts and made read-only or
read-write.
Taking snapshots is fast and efficient. They consume no space initially until data is changed in the
snapshots.
Snapshots are logical copies of the source volumes. They are dependent on the source volumes, and
I/Os to the snapshots might affect I/Os to the source volumes. Limit access to the number of copies of
snapshots created from the same source volume and monitor performance carefully over time to
determine the actual impact to the environment.
Refer to the ME4 Series Administrator's Guide for a detailed description of the features.
3.6.1 Mounting snapshots to recover data
Snapshots are useful to recover data that are deleted, corrupted, or changed unintentionally on the source
volumes. Mount a snapshot created before the unintentional change happened on an alternate host and
retrieve the data using the appropriate tools, such as cp, dd, or application-specific utilities. If data resides on
multiple volumes, the corresponding snapshots of the volumes must be made available to the host.
If the snapshots are mapped to the same host as the source host, additional steps are necessary to mount
the snapshots because the metadata on the snapshot volumes are exactly the same as the source volumes.
These metadata such as disk label, UUID, and volume group metadata, conflict with existing mounted
volumes. It is recommended to mount snapshots on an alternate host instead. However, if the same source
host must be used, use the following procedures to modify the metadata on the snapshots. The snapshot
volumes need to be mapped to the host with read-write access.
1. Use the blkid command to examine the UUIDs on both source and snapshot volumes.
2. If the snapshot volumes contain LVM, change the LVM VG UUID during the import.
3. After mapping the snapshot volumes and scanning the host, identify the snapshot volumes and their
device files (see section 3.1). Use the vgimportclone command to properly change the UUIDs of the
snapshot volume group and physical volumes.
# vgimportclone -basevgname vgme-snapshot /dev/mapper/mpathz