Compellent Live Volume

In another post I made the analogy that Live Volume was similar to storage vmotion from VMware.  Thinking about it a bit more, it’s more similar to VMware Fault Tolerance (but for storage) in that the volumes live in both storage center systems all the time.  The data is kept synchronized, not migrated upon request.  You just switch the “access” from one system to another and the front-end host will follow the storage to the new storage center seamlessly.

Compellent Storage Center 5.0 new features

  • RAID 6 – Not just your mother’s RAID 6.  Compellent has a patent pending on this one as they are mixing the use of RAID 10DP as an initial write destination utilizing their dynamic block architecture that they’ve had all along to mitigate the RAID 6 write penalty
  • Consistency Groups – Yup, what it sounds like….the ability to snap multiple volumes (up to 40 right now) in unison to create cross volume atomic consistency.
  • Storage Monitoring/Alerts – Some simple thresholding that you used to need enterprise manger for, but they made this simple and easy to see so customers can see at a glance how best to plan for additional storage
  • Live Volume – Asynchronous mirroring across storage center systems that provides for seamless host failover between the systems.  Really intended for proactive movement of LUNs to support maintenance activities not so much for DR.  Best used on campus SAN or within a datacenter.
  • Server Mapping – Some pretty cool ways to organize physical and virtual hosts into groups such that when storage is assigned, you don’t have to assign storage to each host individually and/or if you add a host to the group it inherits the storage.  There were a lot of other benefits and integration that took into consideration things like OS version which would warn you about limits you might not be considering that were host dependent and made the system aware of virtual hosts and clusters as they need to be treated a bit differently as we know from a storage perspective.  Makes the storage center more “aware” of this stuff.
  • Virtual Ports – Pretty cool stuff where virtual ports could be transitioned to or combined with other virtual ports thus sharing physical ports.  Allows the system to use less physical ports for redundancy making the system more scalable as less physical ports would be “reserved” for passive failover ports.  Obviously in the event of a failover, you would be sharing physical ports, so bandwidth would need to be considered as two ports worth of traffic would end up on one physical, but you get the idea.