VMware Cloud Community
SeanDA
Enthusiast
Enthusiast
Jump to solution

SRM using different DELL/EMC Clariions

I have a customer who currently has a Dell Clariion CX300. They are looking to replace it with a CX4-120 and would like to send the CX300 to an offsite location for use as a DR solution using SRM.

From my research I believe this solution is possible using EMC RecoverPoint/SE to mirror the two arrays over an IP network, however what is not clear is if I no longer require SnapView and MirrorView (I dont think I do)

Ha anyone done anything similar to this previously, and would care to share their experiences?

Tags (2)
0 Kudos
1 Solution

Accepted Solutions
bladeraptor
VMware Employee
VMware Employee
Jump to solution

Hi Sean

I am writing as an employee of EMC within the company's UK VMware Affinity Team. Please see my answers below in line

"I have a customer who currently has a Dell Clariion CX300. They are looking to replace it with a CX4-120 and would like to send the CX300 to an offsite location for use as a DR solution using SRM."

>>>> This is eminently possible - though as you suggest with the CX300 not supporting MirrorView based SRM (support for this starts with the CX3 family) the only way to achieve SRM with the CX300 and be SRM compliant would be through the use of RecoverPoint as you suggest

From my research I believe this solution is possible using EMC RecoverPoint/SE to mirror the two arrays over an IP network, however what is not clear is if I no longer require SnapView and MirrorView (I don't think I do)

>>>> RecoverPoint uses its own capabilities in conjunction with either a host, fabric based or CLARiiON based write splitter to journal all the nominated writes to a designated SAN 'journal volume'. Neither SnapView or MirrorView are required for this

>>>>The CX3 / CX4 generation of CLARiiON has support for the RecoverPoint CLARiiON splitter allowing those units to split the writes of both VMs with RDMs (which is a limitation when using the host splitter with ESX - in that the host splitter only supports Microsoft VMs using RDMs - hence no SRM support) as well as being able to split the writes of VMFS volumes making it suitable for SRM.

>>>>Fabric based splitting and CLARiiON based splitting are fully integrated with ESX, including support for Site Recovery Manager (as per the VMware SRM comparability Matrix) As the CX300 does not support the CLARiiON Splitter - this leaves you with the option of fabric based splitting from the DR side using something like the Brocade 7600 or the Cisco SSM blade or 9222i Fabric switch My concern with this is that the cost of implementing an intelligent fabric based RecoverPoint solution at the DR side may exceed what the customer will budget for its DR site

>>>> Depending on the profile of the ESX data being replicated to the DR site it may prove cost effective to upgrade the DR unit to another CX4-120 and use MirrorView in conjunction with SRM or if RecoverPoint remains the preferred solution - have CLARiiONs at both the production and DR site capable of supporting the CLARiiON RecoverPoint splitter

Let me know your thoughts

Kind regards

Alex Tanner

View solution in original post

0 Kudos
4 Replies
bladeraptor
VMware Employee
VMware Employee
Jump to solution

Hi Sean

I am writing as an employee of EMC within the company's UK VMware Affinity Team. Please see my answers below in line

"I have a customer who currently has a Dell Clariion CX300. They are looking to replace it with a CX4-120 and would like to send the CX300 to an offsite location for use as a DR solution using SRM."

>>>> This is eminently possible - though as you suggest with the CX300 not supporting MirrorView based SRM (support for this starts with the CX3 family) the only way to achieve SRM with the CX300 and be SRM compliant would be through the use of RecoverPoint as you suggest

From my research I believe this solution is possible using EMC RecoverPoint/SE to mirror the two arrays over an IP network, however what is not clear is if I no longer require SnapView and MirrorView (I don't think I do)

>>>> RecoverPoint uses its own capabilities in conjunction with either a host, fabric based or CLARiiON based write splitter to journal all the nominated writes to a designated SAN 'journal volume'. Neither SnapView or MirrorView are required for this

>>>>The CX3 / CX4 generation of CLARiiON has support for the RecoverPoint CLARiiON splitter allowing those units to split the writes of both VMs with RDMs (which is a limitation when using the host splitter with ESX - in that the host splitter only supports Microsoft VMs using RDMs - hence no SRM support) as well as being able to split the writes of VMFS volumes making it suitable for SRM.

>>>>Fabric based splitting and CLARiiON based splitting are fully integrated with ESX, including support for Site Recovery Manager (as per the VMware SRM comparability Matrix) As the CX300 does not support the CLARiiON Splitter - this leaves you with the option of fabric based splitting from the DR side using something like the Brocade 7600 or the Cisco SSM blade or 9222i Fabric switch My concern with this is that the cost of implementing an intelligent fabric based RecoverPoint solution at the DR side may exceed what the customer will budget for its DR site

>>>> Depending on the profile of the ESX data being replicated to the DR site it may prove cost effective to upgrade the DR unit to another CX4-120 and use MirrorView in conjunction with SRM or if RecoverPoint remains the preferred solution - have CLARiiONs at both the production and DR site capable of supporting the CLARiiON RecoverPoint splitter

Let me know your thoughts

Kind regards

Alex Tanner

0 Kudos
jguerrero
Contributor
Contributor
Jump to solution

Hi Alex

I have other question, I have to apply a change of IP network in the hardware in the secundary site: SAN Clarion EMC CX4 + VC + SRM + 02 ESX.

How to impact this change to my environment?? Can you explain this?

Regards

Jguerrero

0 Kudos
bladeraptor
VMware Employee
VMware Employee
Jump to solution

Hi Jguerrero,

I would like to help you with your query - can you please explain in a bit more detail what you are hoping to achieve on the SR side of your SRM environment.

Clealry one of the joys of SRM is that you can have two completely separate environments, fabrics, storage, hosts, networks and as long as there is storage replication, fibre and IP routing between the two sites you should be able to configure SRM to recover your VMware environment on the remote site

When you say you have to apply a change of of IP network in the hardware in the secundary site - what do you mean by this?

Are you asking about configuring the VMs from the primary site that will be recovered on the secondary site and may, as a consequence, need new IP addresses?

If this is the case I would suggest an SRM guru from VMware would probably do a better job than I explaining how the latest version of SRM has expanded the capability of the solution to help customers provide specifications and scripts to re-ip their VMs on the recovery side

Alternatively if you are talking about the underlying hardwar - storage, ESX hosts, VC on the DR site - these can all be on a separate IP addresses to the production site as long a there is routing for inter site communication and storage replication traffic

Please let me know which of these elements you are focusing on - the IP logistics of recovering production VMs on the remote site or configuring the hardware required for VMware SRM solution on the remote site

Many thanks

Alex Tanner

0 Kudos
jguerrero
Contributor
Contributor
Jump to solution

Hi Alex!!

Is being implemented for a solution of SRM + EMC Clarion CX4-120, the 1st stage is proof: the two environments will be in 1 single site, there are all conducted initial tests as well as the configuration of equipment, then termianda this stage would be carried out the movement of equipment of the second site to another site that has a different network segment, that meant changing the configuration of network: Storage, Host ESX, Virtual Center and SRM.

The question is: How can be done this in a transparent manner, taking into account that it is an environment in which production is going to change...

Regards

Jguerrero

0 Kudos