P4300 Volume Failover & Server Clusters

[rianui]

Hi Gurus,

I have a scenarion where a client has been using old DL servers. Recently he bought BL460c and a c7000 enclosure plus two pairs of P4300 storage (4 nodes). The set up he wants to achieve is that the new servers will be on the primary site using one storage cluster while the other storage cluster will be used on a DR site. The new servers have been set up with about 6 VMs. ESXi resides in the internal disks on the servers but the VMs are running from the storage. How can he achieve failover from the VMs on the primry site to the physical systems on the DR?

Regards,

Rianui

P.S. This thread has been moved from  Storage > Planning to HP StoreVirtual / HP LeftHand Storage. -HP Forum Moderator

[Ravi_K]

Hi

As I understand your query is related to the 4 P4300 nodes.

Customer will need to configure his environment as multisite cluster with 2 nodes on each site to be able to access volumes in event of a site failure. He has to run even number of managers on each site and also configure FOM ( recommended to configure FOM on 3rd physical site).

The above will have to be used in conjunction with VMware Fault Tolerce and Vmotion, I am not a VMware expert so don't know details from the VMware part.

I have also attached a small guide for setting up multisite cluster

Hope it answers your query

Regards

Ravi

[HPstorageTom]

First of all, whether he can use the Multi-Site SAN with the synchronous replication or whether he needs to use the asynchronous Remote Copy functionality of the P4300 systems depend on the available intersite link (ISL) bandwidth and roundtrip latency.

If the roundtrip latency between the P4300 nodes across the ISL is less than 2ms and the ISL bandwidth is at least 1GbE, then he can build a supported multi-site SAN with transparent failover and VMware Metro Storage Cluster (vMSC) support.

This would require as mentioned a Failover Manager (FOM) running on a third site to avoid split-brain scenarios.

The failover of the storage from one site to the other site woud be transparent to the server (i.e. no manual intervention needed). Take a look at the StoreVirtual - VMware - Multi-Site whitepaper for further explanations:

http://h20195.www2.hp.com/v2/GetDocument.aspx?docname=4AA0-4385ENW&doctype=white paper&doclang=EN_US&searchquery=keywords=(OR) p4000 lefthand &cc=us&lc=en

Now, if the roundtrip latency is higher than 2ms or if the available bandwidth is not sufficient for a synchronous replication than an asynchronous replication using the StoreVirtual Remote Copy functionality must be used.

The failover can be made semi-automatic if VMware Site Recovery Manger with the Lefthand SRM Adapter is used. See the LeftHand SRM User Manual for details: http://bizsupport2.austin.hp.com/bc/docs/support/SupportManual/c03687482/c03687482.pdf