cancel
Showing results for 
Search instead for 
Did you mean: 

VCS with 1 physical and 1 vsphere virtual machine

Jeffrey_Ong
Level 3
Partner Accredited
I am not sure if this has been discussed before, i have tried using the seach button. Anyway would like to ask if this is a supported configuration: SFHA for Windows to cluster a physical machine with a virtual machine sitting in a vsphere 4 ESX box. It will be used to failover MSSQL2008 database. Could anyone enlighten me?

If it is supported what are the require pre-reqs? Thank you in advance!

Jeffrey.
1 ACCEPTED SOLUTION

Accepted Solutions

Marianne
Moderator
Moderator
Partner    VIP    Accredited Certified

I don't know if documentation exists that says it is or is not supported. I was actually hoping that someone from Symantec would've responded by now...

My questions with such a setup would be:
Can you guarantee reliable cluster interconnects (heartbeats) ? Remember that the interconnects MUST be independent.
Can you guarantee reliable access from both nodes to the same shared storage?


If you can configure this in a lab environment and do failover tests as well as simulate all sorts of resource failures to test expected behaviour, I don't see a problem with such a configuration. A critical concern in the evaluation process would be to ensure that Single Point Of Failure (SPOF) is identified and eliminated. You can actually do that up front by drawing the configuration and pinpoint any possible SPOF.

View solution in original post

4 REPLIES 4

Marianne
Moderator
Moderator
Partner    VIP    Accredited Certified

I don't know if documentation exists that says it is or is not supported. I was actually hoping that someone from Symantec would've responded by now...

My questions with such a setup would be:
Can you guarantee reliable cluster interconnects (heartbeats) ? Remember that the interconnects MUST be independent.
Can you guarantee reliable access from both nodes to the same shared storage?


If you can configure this in a lab environment and do failover tests as well as simulate all sorts of resource failures to test expected behaviour, I don't see a problem with such a configuration. A critical concern in the evaluation process would be to ensure that Single Point Of Failure (SPOF) is identified and eliminated. You can actually do that up front by drawing the configuration and pinpoint any possible SPOF.

Jeffrey_Ong
Level 3
Partner Accredited

Hi Marianne,

Thank you very much for your input, really appreciate it. Seems like there isn't any response from Symantec as well up till now. As for the heartbeat, it will be dedicated NICs assigned to the Virtual machines so i guess this will eliminate the heartbeat and any network latency issues.

As for the access to the shared storage, i am thinking if vmware's Raw Device mapping of a LUN is supported by SF? I don't have a lab rite now as i am actually proposing a solution at the moment.

Marianne
Moderator
Moderator
Partner    VIP    Accredited Certified

Hi Jeffrey

Maybe contact your local Symantec SE? He/she should be able to contact a Symantec product specialist who will be able to give best possible advice.

Wally_Heim
Level 6
Employee
Hi Jeffrey,

Marianne is right on track with this one.  You need to have heartbeat networks that are in common between the physical and virtual nodes and you need to have access to the same shared storage (or replicated storage between the nodes.)

If you are able to do both of those items then the cluster would be configured as normal.

SFW-HA/VCS supports clustering Physical to Physical, Physical to Virtual or Virutal to Virtual.  It all depends on how you can configure your network and storage components.

Thanks,
Wally