Your Browser is Out of Date

Nytro.ai uses technology that works best in other browsers.
For a full experience use one of the browsers below
Dell EMC PowerMax and VMAX All Flash: TimeFinder SnapVX Local Replication
Executive summary Introduction TimeFinder SnapVX overview
Creating snapshots
Generation numbers
Time-to-live value
Host writes to source
Redirect-on-write
Replication cache
Shared allocations
Linking target volumes to snapshots
Relinking
Unlinking
Defining state
Linking a larger target
Source-volume or linked-target SRP relationships
Restore operations
Cascading snapshots
Reserved capacity
Interoperability
Storage group operations
Secure snaps
Online device expansion
Role Based Access Controls
Implementing TimeFinder SnapVX PowerMaxOS Q3 2020 release enhancements TimeFinder legacy modes: Mirror, Clone, and VP Snap Conclusion Appendix A: TimeFinder SnapVX state table Appendix B: Geometry compatibility mode Appendix C: Monitoring the copy process Appendix D: Nocopy Linked Target behavior before HYPERMAX OS 5977.810.184 Appendix E: Scheduling snapshots with Unisphere for PowerMax Appendix F: System resource planning for capacity and performance Appendix G: Technical support and resources

Cascading snapshots

Home > Storage > PowerMax and VMAX > Data Protection > Dell EMC PowerMax and VMAX All Flash: TimeFinder SnapVX Local Replication > Cascading snapshots

Cascading snapshots

  • Cascading refers to taking a snapshot of a linked target and linking a target to a snapshot of a linked target. All tracks on a target volume must be defined before a snapshot can be taken of the target volume.

    The following figure illustrates one example of cascaded SnapVX sessions.

    Figure 11.    Cascading snapshots and linked targets

    The cascading capabilities of SnapVX are very robust. However, several traditional use cases for cascading clone or snap sessions no longer exist with SnapVX.

    Since point-in-time data for SnapVX snapshots is stored in the SRP and cannot be directly accessed by the host, the point-in-time of the snapshot is always pristine, and there is no need to make a “Gold Copy” clone to cascade from. Even if a target is linked to the snapshot and written to, the point-in-time of the snapshot does not change. This reduces the number of full volume copies that are required in most use cases making SnapVX very space efficient.

    Furthermore, the defining mechanism allows reads of tracks that do not reside on the target volume to be serviced directly from the SRP and not redirected to the source volume as would be the case with traditional snaps or clones. This not only improves performance of the read but also reduces the possibility of target reads affecting performance of the source volume.

    Cascaded SnapVX sessions are supported with the following considerations:

    • There is no limit on the number of cascaded hops that are allowed.
    • Beginning with PowerMaxOS 5978, Solutions Enabler 9.0 and Unisphere 9.0 users can take a new snapshot of a linked target that has the define process still in progress. Earlier code versions only allowed a snapshot of a linked target that is in either the defined or copied state.
    • A cascaded snapshot can only restore to a linked target that is in copy mode and has fully copied.

    Unlinking a fully defined nocopy target rather than changing to copy mode will also allow the snapshot to be restored.

    • The cascaded chain can be broken at any point by unlinking the target from the snapshot with the following rules:
    • A snapshot that has a linked target cannot be terminated.
    • Prior to HYPERMAX OS 5977.810.184 a linked target that has a cascaded snapshot must be in copy mode and fully copied before it can be unlinked. This is not required with HYPERMAX OS 5977.810.184 or later.