Skip to main content
Start a Conversation

This post is more than 5 years old

Solved!

Go to Solution

satyaranjan1

1 Rookie

 • 

31 Posts

5439

December 18th, 2015 06:00

Recoverpoint Replication

Hi Experts,

IHAC who is running his setup in a 3rd Party owned DC for the past 4 years. They have decided to set up their own data center about 70 miles from their present one. Existing Data is on 3rd Party Storage ( IBM DS8700). Majority of his compute estate is virtualised with just a couple of Physical servers. At present all of this is provided and managed by the 3rd party.

The issue we are facing is with the migration of data. At present there are no links available between the Present DC and the DC the customer wants to move to. Link with very small bandwidth will be available after a couple of weeks, however customer has a deadline to finish the migration. He wants us to finish a copy of the data to the new storage ( Data is a little in excess of 200TB, Proposed array is a VNX5800), move it to the new data center ( lift and shift) and sync up with the differential delta as he plans to keep the production running during the time of the lift and shift at the 3rd Part DC.

We were contemplating of putting a VPLEX and RP( Cluster A = 2 X Gen5 RPA), to virtualize the DS8700 and replicate the Data to the VNX5800 + RP ( Cluster B = 2 X Gen5 RPA). This is going to be within the same DC. Once the initial copy is done, Pause the replication, lift the VNX5800 with the RP to the Target DC when the links are up ( estimated time is 2-3 days) resync the delta.

Need your suggestion in this regard, based on your expirience is the above mentioned plan feasible Technically or based on the feature set of RP is there a way we can have a delta only resync and not the whole copy once we lift/Shift the VNX+RP .

Thanks in advance

forshr

2 Intern

 • 

1.1K Posts

December 21st, 2015 09:00

The correct option is to use the 'clear_markers' CLI command as follows:

  1. Enable the CG with no replication or transfer – enable_group group=xxxxxx start_transfer=no
  2. Clear the markers for the CG copies which forces RP to avoid the full sweep on the basis that due to the sync replication mode the source and replica volumes are identical – clear_markers group=xxxxxx
  3. Start replication by starting the transfer – start_transfer group=xxxxxx
  4. The group will move to an Active state immediately but avoid a full sweep. You may experience a short init.

Regards,

Rich Forshaw

RecoverPoint Corporate Systems Engineering

fadliz

153 Posts

December 18th, 2015 06:00

Is there a link bw the prod site and the new DR? If so, take a look at the DC migration scenarios in this WP.

https://community.emc.com/docs/DOC-40044

Zahid Fadli

EMC2

Twitter: @zDataGuy

paul_montoya1

5 Posts

March 26th, 2016 10:00

Zerto would work well for the VMs.  For the physicals attached to the IBM storage that could get a little tricky.  If the customer could take an outage on the IBM side, you could use pass through disks with datacore and then replicate those site to site to a Datacore/VNX on the target side.  During cutover, you could remove the target datacore passthrough luns leaving the hosts attached to only the VNX 5800.

Was this post helpful?

View All

No Events found!

Top