1

I'm looking into various options for setting up massive high-availability storage in Windows Azure, to use together with my Azure VMs. The VMs need to be able to work with the storage as if were a file system. Let's say that I have three VM's running in Azure, that all need to share the same storage. It is important that the shared storage is highly available, with minimum failover time. Add to this that I need massive storage - we are talking tens of terabyte, and maybe hundreds in the long run.

I'm currently consider setting up Distributed File Sharing (DFS) on multiple Azure VM instances, to create a multi-instance file server environment, that my other VMs then can utilize. I've found a good resource on how get that running. This solution looks like a good setup, but it comes with two caveats:

  1. From what I can understand, the failover time to switch to a redundant file server instance would be somewhere between 60-90 seconds. Given that the file server instances would be restarted a few times a month as part of regular OS updates etc., I believe we would still be able to keep a 99.95% SLA, but it would be preferable with a shorter failover time.

  2. Given Azure's current limitation on how many VHDs you can attach to a single VM (currently 16 x 1TB VHDs for large VM instances, and the fact that I need redundancy, I would have to spin up 2 new file server VMs for every 16TB storage I need to add.

Is there a more cost efficient way to create massive high-availability storage in Azure, given that the VM's that utilize the storage need to work with the storage as if it were a file system? And if not, is there a way to shorten the failover time of this setup even futher?

1
  • Doesn't sound like a too bad approach. You could script most of the stuff using PowerShell / Azure API's, I guess
    – MichelZ
    Apr 23, 2014 at 12:18

0

You must log in to answer this question.

Browse other questions tagged .