Moving Dedup Store with little downtime
I encounter a problem with Windows Block size.
My Dedupe store would not be able to be larger than 16TB, so we need to format volume again with a larger block size
To solve this, we consider to
1) Add another storage array on another raid controller
2) robocopy dedup store to this new storage array (while backupexec still running backup) (would take a lot of time maybe one day at least)
3) Stop all backupexec services, then do another differential robocopy (would take a few minutes, or maybe one or two hours)
4) target new volume for dedup store in backupexec to use the new array
5) format main array with a larger block size and do robocopy in the other way
6) Stop all backupexec services, then do another differential robocopy
Would it work and should I be able to get my deduplication store up ?
The best way that I've found is to stop all services, copy directory, make sure new location has same path/drive-letter, start services. It works.
Using rsync/robocopy and resyncing at a later time is an unsupported complexity that is not needed when a proper outage is planned.
Moving a number of TB's of data over the local PCI buss and fast disk is not terribly slow. It's better on a SAN with snapshots and clones, but workable with a quality JBOD and SAS too.
