UNSOLVED
Swing in a new Celerra? Replication Guru's???
Question for the Replication guru's out there...
Here is my CURRENT environment.
SOURCE CELERRA - NS40 (CEL-A) - All production is currently running here..
CEL-A replicates to local Celerra (CEL-B)
CEL-A also replicates to an offsite array (CEL-X)
CEL-A ----> CEL-B
|
|
V
CEL-X
We have two new Celerra's that I'm in the process of replicating the VDM's and File Systems with the intentions of bringing those resources online.
We'll call them (CEL-1) and (CEL-2)
STEP 1 Create new replications to CEL-B and maintain the replications after the Celerra Switch Overs from CEL-A to CEL-1 and CEL-2.
CEL-1 ---> C C
E + E ------------> CEL-X
L L
CEL-2 ---> B A
STEP 2
My intention is to get CEL-A out of production and swing in CEL-B as the target for local replication from CEL-1 and CEL-2. CEL-B will then replicate offsite to CEL-X.
CEL-1 ---> C
E ------------> CEL-X
L
CEL-2 ---> B
Can I create new replications from a different Celerra and keep the file systems intact on CEL-X without having to re-replicate everything?
Rainer_EMC
6 Operator
•
8645 Posts
1012
0
Posted September 6th, 2011 06:00
Should be in the 6.0 Replicator manual since it's already available there
I just can't check right now
Rainer_EMC
6 Operator
•
8645 Posts
1012
0
Posted July 19th, 2013 08:00
Please download the latest Replicator manual and search for „common base checkpoints” there
umichklewis
4 Apprentice
•
1190 Posts
1016
0
Posted July 9th, 2015 08:00
What Rainer is referring to in the document is "Common Base Checkpoints". Refreshing a replication with common base checkpoints between multiple Celerras allows you maintain the "state" of filesystem and VDM replications, using a user-created checkpoint that you create as the "base". In your example, when you're ready to replace Cel-A -> Cel-X replication with Cel-B, you would create common checkpoints with the "fs_ckpt" command on Cel-B and Cel-X, then use "nas_replicate -refresh" against those checkpoints on each Celerra. This makes a common "starting point" for the new replication you're creating between Cel-B -> Cel-X. This way, you avoid having to re-replicate the VDMs and filesystems from scratch.
This is pretty easy to test out on the VNX simulator for Proof-of-Concept or creating documentation.
Let us know if that helps!
Rainer_EMC
6 Operator
•
8645 Posts
1016
0
Posted July 9th, 2015 14:00
yep - thats what I meant
Its pretty simple
you create a user checkpoint on each of the systems and then use nas_replicate refresh to get them all to the same state
when you later a replication between them and create a new replication the system will automatically determine that these checkpoints are suitable as a common base and will just do in incremental update instead of a full copy.
for example if you have
A -> WAN -> B (to be replaced)
you create a RevpV2 session to the new system
A -> WAN -> B (to be replaced) -> C (new system)
then create user checkpoints and A,B,C and refresh them
Now you can remove both replications and setup a new one like this:
A -> WAN -> C (new system)
without incurring a full copy
hope that helps
Rainer
P.S.: another use case for refreshing checkpoints over a replication session would be create application constistant state on a src system, refresh the local cpkt, refresh local to remote cpkt and now you have the exact same content there for backup purposes
Rainer_EMC
6 Operator
•
8645 Posts
1012
0
Posted September 6th, 2011 06:00
Just wait a little bit for the next release with Replicator incremental attach