Backup Exec 2010 Disaster Recovery of deduplicated storage folders

[kpapreck]

I am trying to put together an environment with Backup Exec 2010 deduplication storage folders but using SAN based replication.  Under the admin guide for disaster recovery of deduplication storage folders it talks about taking a snapshot of the data in which you can send to tape.  In this DR to tape, it includes the folder, the contents of the folder, and the database of the folder.  For this, is it then backing up the deduplicated data to tape in a deduplicated format? 

Are there any issues of running backup exec 2010 deduplication storage folders and using a non OST device to replicate the data?  Will I just need to perform a re-catalog to restore the data at the DR site for it to see all of the deduplicated backup sets?

[teiva-boy]

 When exporting deduped data to tape, the data is re-assembled into it's FULL data set, and written to tape as the original data file.  You cannot send deduped data to tape in its reduced size.  

The benefits to this is that, when you need to restore, it's a simple operation, just pull the file from the tape.  If it were deduped, the restore would take somewhere between 10-50x longer as multiple tapes need to be searched for bits and pieces of the deduped data to reassemble it.  In some other products, the entire tape needs to be restored to disk first as a staging area..  What if it were TB's of data?

We backup so we can restore, not so that we can backup...

I think what you want in your last paragraph is the CASO option with distributed catalogs, or centralized catalogs on the CASO.  You can duplicate backups from media server to media server, but media server #2, doesnt have any catalog data of the backup.  CASO would solve that issue.

For SAN Based replication, use that for your quick recoveries that you can do from the snapshots themselves.  Use the media server duplication feature to replicate backups to another site.  I've yet to meet a single replication product that can do it all.... And this case is no different.  SAN based for some data types, Exchange CCR/LCR, SQL log shipping, media server duplicate copies, etc...  There is no magic bullet.. Yet.

[kpapreck]

THe issue i see with using CASO and duplicate backups is that you will lose GRT for Hyper-V and SharePoint.  This is the reason for planning to go down the SAN based replication route.  Does anyone have any suggestions?

[teiva-boy]

Use BE to backup to a PureDisk Server/Storage Pool in Site A
You replicate the PureDisk data to another PureDisk server in site B. (along with a corresponding BE server there too)

CASO is used to be catalog aware of Site B and to control the replication.

When you do a restore, BE will retrieve the data from the PureDisk Storage pool and re-hydrate it.  


There is the workaround for the GRT piece that BE doesnt have natively as far as I can tell...  Or you buy a Quantum or Exagrid Dedupe appliance for each site instead of PureDisk.