StoreVirtual Storage
1752782 Members
5903 Online
108789 Solutions
New Discussion

Re: Round Robin or Failover (Vendor Specific)

 
TLH3385
Occasional Visitor

Round Robin or Failover (Vendor Specific)

Hello,

 

I am setting up a new HyperV cluster with some new SAN volumes and I am wondering what is the best option to choose when connecting to a P4000 via iSCSI. Currently the systems are set to vendor specific but I am looking to get more performance from this cluster.

 

It seems like since Round Robin uses active/active, I should get double the bandwidth and also the system should still run if one of the NICs fail. Is there any downside to this method?

 

Any suggestions or best practices?

 

I have:

2x P4000

6x Dell Poweredge servers; each server has 2x intel 1GB nics dedicated to iSCSI traffic

Windows 2008 R2 Hyper-V (Full install)

 

Thanks,


Trent

4 REPLIES 4
oikjn
Honored Contributor

Re: Round Robin or Failover (Vendor Specific)

HP manual says BP is round robin.

TLH3385
Occasional Visitor

Re: Round Robin or Failover (Vendor Specific)

Do you have a link to that except or manual?

oikjn
Honored Contributor

Re: Round Robin or Failover (Vendor Specific)

read the P4000 Windows Solution Pack user guide.

 

That entire document is helpful and will answer most of your questions, but the detail you are looking for is on page 19.

Emilo
Trusted Contributor

Re: Round Robin or Failover (Vendor Specific)

I am not sure how many nodes you have in the cluster but DSM/MPIO should really help with your performance and provide fault tolerance for your iSCSI host. You want to use Round Robin if you can , it can be kind of tricky to setup properly but well worth it.

 

HP P4000 DSM for MPIO Deployment Guide

http://bizsupport1.austin.hp.com/bc/docs/support/SupportManual/c03041928/c03041928.pdf

 

Supported load balancing options
Only two load balancing options are supported, Fail Over and Round Robin. For the Microsoft DSM
with Windows 2003 Server, only the Fail Over Only option is supported.
• Fail Over Only – Also called active/passive MPIO. Two (or more) I/O paths are built between
the server and the storage. One path is actively used for I/O to the storage. The other paths are
available for failover only in the event the primary path goes down.
• Round Robin – Also called active/active MPIO. Two (or more) I/O paths are built between the
server and the storage. All paths are actively used for I/O to the storage.
In the HP DSM for MPIO, Vendor Specific is selected by default. This option is the same as selecting
Fail Over Only and will not be available after you select another option. In the Microsoft DSM, Fail
Over Only is the default selection for all Windows versions, except for Windows 2008 R2, for which
Round Robin is the default.

 

Hope this helps please mark at as solved if it does