A couple of days ago 3PAR made more announcements about their added features in the upcoming major release of their InServe operating system for their enterprise fibre channel and iSCSI storage arrays. If you missed their previous announcements, click on 3PAR on the right side of my blog for the other articles.
Of the three features announced recently, I’ll focus on autonomic groups since that is a feature my project will certainly be using. If you working in a highly virtualized environment with dozens of ESX hosts, then you know how complex storage provisioning can become. In an ESX cluster all servers must see the same set of LUNs on the storage array. This enables features such as vMotion, DRS, and HA.
But the problem becomes in presenting storage to several clusters and dozens or hundreds of hosts. Typically in a storage array with LUN masking you would need to individually present a LUN to each and every ESX host. If you forgot one host, then vMotion could crash and bomb. Or worse yet, if you present a LUN to the wrong server all of the data could be corrupted or a security breach occur. Human error in any environment, and particularly large environments, is a big problem.
What has 3PAR done about this? They have developed autonomic groups. An autonomic group is a collection of hosts and LUNs. Adding a new ESX host to an existing cluster with 30 LUNs? With autonomic groups now you just need to add the ESX host to a single autonomic group and a couple of clicks later all LUNs are presented to the new host. Or, if you have an existing ESX cluster and you add a new shared LUN, just add the LUN to the autonomic group and its exported to all ESX hosts with a couple of clicks. Massive reduction in the chances of human error, not to mention a significant reduction in provisioning time.
In addition to simple storage and host provisioning, they’ve also tied in snapshots. So you could snapshot an entire ESX cluster and the associated LUNs with a couple of clicks. Now all 3PAR needs to do is provide Windows VSS support for *ALL* ESX VMs on the LUNs in the autonomic group so that hardware based snapshots are 100% transactionally consistent when the autonomic group is snapshot. Let’s hope that will be in their Recovery manager for VMware product.
The other products they announced in the 3PAR scheduler, which is a scheduling engine to automate routine tasks. And finally, host explorer helps automate the detection of WWNs and other host based information.