Data ONTAP Discussions

Dedupe with 7.3.1 and Snapmirror - best practices

Are there any published best practices running dedupe with 7.3.1 and VSM?

With metadata now being stored in the aggregate but replication (at least in my case) occuring at the volume level, what are the recommendations for dedupe operations on the DR Filer if the need arises to break the mirror and make the DR writable?

For one I'm guessing there's no metadata present as its all hiding at the aggregate level on the production system so should you recreate the metadata with a 'sis start -s'?

Do the same processes need to considered when reversing replication to restore production?

Is this process more straight forward if leveraging Protection and Provisioning Manager 3.8?




Re: Dedupe with 7.3.1 and Snapmirror - best practices

TR3505 covers many of the best practices for dedupe, including how it integrates with VSM, so that's well worth a read.

Basically with 7.3.x when you failover the dedupe metadata isn't there, but the old data will still be deduped.  Any new data will be deduped with any other dedupe data, but you need to run sis start -s to re-create all the dedupe metadata.  This is explained more clearly in TR3505.