Server Fault is a question and answer site for system and network administrators. Join them; it only takes a minute:

Sign up
Here's how it works:
  1. Anybody can ask a question
  2. Anybody can answer
  3. The best answers are voted up and rise to the top

I'm currently evaluating a Win2k8 two node file server cluster that is connected to a Win2k8 Storage Server iSCSI SAN. The failover is working perfectly, but I'm now wondering what will happen if the iSCSI SAN box falls over. Is there a way to make the SAN highly available? I was thinking that using dfs would help replicate the data to another SAN, but I'm not sure if there was a way of automatically failing over from one SAN to another.

Many thanks,


share|improve this question
up vote 1 down vote accepted

Because Win2k8 Storage Server is not a 'dual' controller, cache coherent SAN device, a secondary system would be needed to provide redundancy. I don't know if Win2k8 Storage Server offers any built in cluster features that could accomplish this or if a product like Double-Take would better accomplish the end goal.


share|improve this answer
Thanks Jeff, I think I might switch from Microsoft's iSCSI target to StarWind as it comes with its own iSCSI-based replication engine. I suppose it's possible to write a script that will swap where the iSCSI initiators are pointing if an iSCSI target has been detected as being out of service, thus having iSCSI SAN redundancy. – Andrew May 28 '09 at 13:16
+1 for pointing out that the cache mirroring/coherence across the SAN heads and/or controllers is actually the feature that you pay for. – pfo Dec 14 '10 at 13:45

You can always replicate the SAN data indeed but considering the price and specs of SAN boxes, there are highly available to themselves. This is only true if you actually get a decent SAN obviously.

This does not mean that your disks become more reliable though... Always use some sort of RAID technology to protect (i.e. not RAID0) your data accross disks.

share|improve this answer
Thanks for your help Antoine, I meant more along the lines of the SAN is still a single point of failure, even though there is a cluster of file server boxes. Is there away to have another SAN that will become the main file store automatically if the first SAN goes down? Many thanks, Andrew – Andrew May 28 '09 at 9:24
Yes you can. You can use DRBD for block level synchronization and switch in between the two SANs if one goes down. – Antoine Benkemoun May 28 '09 at 13:14

Typically you don't replicate the entire SAN - you replicate the switch to provide an alternative router to the disk arrays, and ensure the disks are in a RAID configuration.

I couldn't find my favourite link explaining it, but Google went one better: IBM's redbook on SAN redundancy.

share|improve this answer

Take a look at the LeftHand SAN solutions from HP (formerly from LeftHand Networks). They provide a highly available iSCSI SAN product. We've been using it for nearly 2 years as a block storage for everything in our office. Works very well and the HA failover generally works as advertised, provided your initiator is good.

share|improve this answer

NetApp also provides a clustered highly available iSCSI system like the 3020HA

share|improve this answer

I'm surely know that StarWind can provide High Availability with automated failover and failback.

There are also LeftHand, but it isn't too cheap, you know ;)

Openfiler have HA, but it is active-pasive.

share|improve this answer

You could use the software from Datacore to provide the SAN, it supports real time replication of the data between 2 locations in a synchronous manner.

share|improve this answer

In our company we've configured the active-passive cluster using StarWind 5.0 HA. What can I say? Seems to be working okay. The pro of such approach is that we had no need to buy special hardware but turned into SAN a usual Intel server. Here is more information about this stuff:

share|improve this answer

Your Answer


By posting your answer, you agree to the privacy policy and terms of service.