HPE Storage Users Group

Hi all

1st what a great site and even though this is my 1st post I have gained some fantastic knowledge from here.

So I have 2 EVA8400 using CA (split sites connected via dark fibre) and we are moving to 3PAR 7400's (4 nodes).

As we know CLX we have upgraded our EVA licences and will still be using CLX until we feel comfortable with Peer-Persistence.

What I can't find is a procedure to migrate the hosts in a split cluster using CLX from the EVA's to the 3PAR's, I'm fine with the online migrations but need to get it straight in my head how to proceed in this scenario.

The EVA's are at the latest firmware and command view is also latest version.

Many thanks in advance.

Graham

I'd suggest you call in the experts, CLX is rock solid once configured correctly but can be a bit of a sensitive beast around the setup etc.

So I know CLX that's not my query, installed multiple times, my query is the process or if any migration documentation is available (unable to find any to press).

My thoughts would be the following but just wanted to see if anyone had done it before.

1. stop the cluster resources
2. begin the disk migration from within command view (offline)
3. change the zoning.
4. uninstall clx for EVA, install clx for 3PAR, Setup the multipath for 3PAR.
5. replicate the data from production to DR (so that the volumes are available for the host @ DR) using remote copy.
6. reboot the server (with the cluster service set to manual)
7. reconfigure the cluster to have the new resource (3Par clx)

anyone any views ?

Graham

Hi all and Merry Xmas.

Just wanted to post a reply while it was fresh in my head - I contacted HP about this and it seems that this hasn't been done or nobody has written the steps required.

Well I am glad to say that I successfully migrated a stretched cluster running on the EVA's using CLX for EVA to the new 3Par's using CLX for 3Par and here is my write up of the steps I used.

With cluster up and running
1. Install 3Par CLI.
2. Install CLX for 3PAR.
3. Setup MPIO for 3Par on said host in preparation.
4. If using Bi-Directional replication (Application running on both sites with 2 Data Replication Groups) fail all over the one site so that all resources are running from a single server at a single site.
5. prepare zoning
a. Single port zoning (Single EVA port on controller A. to Single 3Par port on controller 0 and another single EVA port on controller B. to another single 3Par port on controller 1 (I chose one from each fabric to give resilience)) and enable for online migration using command view EVA.
b. Host to 3Par (Production nodes to Production 3Par, DR nodes to DR 3Par (as many ports as required when live (multiple HBA’s/fabrics/controllers)) don’t activate yet.
6. Setup ports zoned in 5a. to be peer ports and that the EVA is shown as connected.
7. Stop cluster services at the DR site.
8. Remove the CLX for EVA resource(s).
9. Remove Data Replications Group(s).
10. Initiate migration on the command view server of all drives presented to clustered host.
11. When required activate prepared zoning in 5b. (I chose to reboot the server at this point so that the applications didn't have issues)
12. When clustered node is rebooted is should pick up the drives it had configured but these will be on the 3Par now.
13. Setup Remote Copy groups same/similar to what was setup on the EVA (Data Replication Groups).
14. Setup DR host access to DR 3Par. (I also changed the host persona in production from 2)
15. Watch for data migration to finish.
16. Export volumes on DR 3Par.
17. Setup CLX to access the 3Par on clustered nodes.
18. Setup CLX resource(s) using CLX for 3Par as was removed using CLX for EVA in step 9.
19. Check volume access from DR Host to DR 3Par.
20. Start cluster services at the DR site.
21. TEST FAILOVER.

I hope this helps others out.