VMWare with 2 SAN, best redundant strategy


Im going to setup a VMWare farm with 4 VMWare Enterprise servers. I have 4 physical servers and 2 physical SANs.
My question is:

- Is it possible to make VM machines runs on 2 SANs simultaneously ? If 1 SAN failed then the other SAN will take over?
- What is best redundant strategy do you suggest to utilize 2 physical SANs ?

Thank you.
Who is Participating?
za_mkhConnect With a Mentor Commented:
No you cannot have a VM run off two SANs simultaneously. What you can do is have snapshot your VMFS volume(s) from SAN(1) to SAN(2) to have a backup copy of your VM's. However, the point of VMWare ESX with SAN is that you can use features like VMOTION and HA to ensure that your VM's are always up and running if a host fails. If your SAN fails, then other options need to looked at.
In this case, you could use VMWare vCenter Site Recovery Manager to 'fail over' to your second SAN since I would class a failure of a SAN as a Disaster! http://www.vmware.com/products/srm/
But really this does depend on what type of SANS you have and your infrastructure.
Exactly what I was going to say...user your extra hardware for a test environemtn or DR/BC.
acsmedicConnect With a Mentor Commented:
On the surface this is an easy question, however when you start to think about it you are looking at a complex thing. SRM is the way to go in this situation as long as your SAN's are able to replicate to one another AND your SAN's are supported by SRM's SRA (Storage Recovery Adapter) http://www.vmware.com/pdf/srm_storage_partners.pdf

Really the question is what is your goal? Utilize both SAN's in production due to space IO concerns? or provide failover and redundency? As long as space and IO are not tight SRM as noted above is the right answer.
Introducing Cloud Class® training courses

Tech changes fast. You can learn faster. That’s why we’re bringing professional training courses to Experts Exchange. With a subscription, you can access all the Cloud Class® courses to expand your education, prep for certifications, and get top-notch instructions.

yowcsaAuthor Commented:
We have 2 buildings (500m distance)  and 2 DELL CX4 SANs, my main goal is if 1 SANs suddently die the other SANs will automatically take over and the systems keep on running.

What is your suggestion for best redundant strategy to utilize 2 physical SANs ?
za_mkhConnect With a Mentor Commented:
Well we have a similar setup where we have two SAN's in two buildings.
Each one data volumes is replicated to the other. Our solution is nearly fully redundant i.e. two Fabrics, etc, etc. The configuration is such that there is an active Path (ESX Server A connected to SAN A in Building A) but it also has a redundant path to the same LUN's etc being hosted on SAN B in Building B via Fabric B so if SAN A fails, it will automatically fail over to SAN B.
Of course both SAN A and SAN B are replicating to each other. You will need to consult you SAN documentation or your SAN Vendor on how you would configure that.
So yes it is possible in that scenario without needing SRM, etc.
500m is not so far away and it would be possible to use setup a mirror with some SAN virtualization software in front in order to automatically use the live site and sync back once it's up again.

Solutions that can do this include DataCore SanSymphony/SanMelody, IBM SVC, EMC InVista, FalconStor &  LeftHand.


Datacore sanmelody does this.  We have three Esx boxes and two datacore sanmelody in synchronous MPIO via ISCSI.  One Datacore San is considered the primary, the other the secondary.  If either SanMelody server goes down the other keeps on going.  Though if the primary goes down unexpectedly sql servers and exchange complain for about a second.  We have two buildings about 2 miles apart, we put 10g  fiber between the two.  We placed one of the ESX boxes and one of the datacore sans in the other building and vlan'ed the iscsi over the fiber.  We can vmotion between all three esx boxes /both buildings and datacore sanmelody works fine synchronous between both buildings.  In my tests it took longer to drive to the other building then it did to get things back online.  We also setup a asynchronous datacore server in a third building for extra backup
Question has a verified solution.

Are you are experiencing a similar issue? Get a personalized answer when you ask a related question.

Have a better answer? Share it in a comment.

All Courses

From novice to tech pro — start learning today.