Vsphere storage redundancy

B

bds1904

Gawd
Joined
Aug 10, 2011
Messages
1,007
What are some options for storage redundancy when using vsphere?

Right now I have a 3 host cluster running HA and DRS using shared iSCSI storage. What are my options as far as redundant storage? My current storage setup is a freenas box running ZFS and can handle a drive failure, but I am interested in being able to handle a complete storage server failure.

Options?
 
4. VSA cluster running on your hosts using local storage
5. Cluster based iSCSI solution like Lefthand, Starwind HA

No matter what you choose, option 1, backing up your VMs, needs to be top priority, then look into how to make your array redundant.
 
Last edited:
Child of Wonder said:
4. VSA cluster running on your hosts using local storage
5. Cluster based iSCSI solution like Lefthand, Starwind HA

No matter what you choose, option 1, backing up your VMs, needs to be top priority, then look into how to make your array redundant.
Click to expand...

Agreed, are there any automated backup features in Vsphere?
 
If you're under support and upgrade to 5.1 you should get VMware Data Recovery, though it's reported to be pretty damned awful and just about the last thing you want to be counting on when you need your data back.

vSphere Replication might also be worth a look though that's more about having a recent replica than about backup in the traditional sense.

To handle a "storage brick" level failure you need more than a single brick so VSA such as HP P4000 or VMware's own VSA are probably your two main off the shelf options unless you want to start rolling your own.
 
When you say redundant storage, what do you mean? Do you mean you want to fail over easily should your iSCSI storage goes down? For that you could do replication with the builtin vSphere replication or other tools. Or you can do backups which are simpler, but take longer to recover.

vSphere 5.1 now has VMware Data Protection, which is different than the older VDR. It works..though I've had some quirks with it.
 
I am running a Lefthand cluster for my ESXi boxes. All of the data sits on an Equallogic cluster.
 
you need to check your rto and rtp requirements

if you can live with some downtime and can afford to lose some data - vm replication and vm backups should be enough (running vms from local storage is fast)

if you cannot afford downtime (ticket booking office, mrt scanner, police camera storage etc) you absolutelu need 3) storage sync replication

geiger said:
1.VM Backups
2.VM Based replication
3.Storage based replication
Click to expand...
 
+1

running raid (and sync lun mirror is a netowrk raid basically) does not eliminate a need in backup

Child of Wonder said:
4. VSA cluster running on your hosts using local storage
5. Cluster based iSCSI solution like Lefthand, Starwind HA

No matter what you choose, option 1, backing up your VMs, needs to be top priority, then look into how to make your array redundant.
Click to expand...
 
You must log in or register to reply here.
Back
Top