2012-10-05 12:40 PM
We are getting ready to layout hardware upgrades using data migrations and want to plan how best to accomplish this task.
The night of the cutover, FilerA_new becomes FilerA, FilerB_new becomes FilerB, current FilerA becomes FilerA_new(to be decomissioned), and so on.
With protection manager we can add the aggrs to the existing system resource pools and migrate using PM. However we are worried that this migration process will cause issues when the system names are all swapped, as 8 systems all swap names(and all 8 names are in PM managing the same volumes.)
Now the long and arduous way would be to ensure the reaper process is disabled, relinquish all the primary dirs, manually update everything involved, then import all the relationships. This isn't entirely practical as there are 1000+ volumes involved. It could be scripted, but the amount of time and manual checking will be problematic.
Is there a process using the PM migration that we can follow to ensure that all of the name swaps do not appear as new relationships and attempt to baseline everything?
2012-10-15 07:02 AM
If its just head swap, then you dont have to go through the arduous process of PM, but if you are going to manally VSM the data from FilerA to FilerA_new then yes you will have to go though the pains.
I am attaching a doc that I wrote on how to keep PM dataset happy when primary is migrated. Hope it helps you.