Protecting INDEX Locations

I am working on a very large installation of EV for archiving and Journaling EMAIL - we expect in excess of 150,000 accounts over the next few years.

We have been advised by the BA's that FULL Indexing is required for both user and journal archiving.

My estimates suggest that we could end up with in excess of 100 TB of Index store over time and it will continue to grow.

This presents us with a significant pressure on backup using contemporary methods in our current backup environment.

Can anyone comment on using Array based replication of indexes instead of backup.

We are going to use Veritac Clustering with EMC Mirrorview for the EV servers (using the appropriate Agent to integrate with Veritas)  and the indexes will be rpelicated to a geographically distant site.

I would like to find some evidence that this would/could be sufficient protection for indexes for DR.
Any comments feedback would be most appreciated.

4 Replies

Hi Yes, you should be able to

Hi Yes, you should be able to do this array based, as EV is able to update indexes once they are restored. If you want similar functionality for your vault stores, have a look at http://www.evtools.net Cheers

Quadrotech - EV Migration and Office 365 Solutions, creators of Archive Shuttle.

regarding backups we are

regarding backups we are looking into the option of an EMC Avamar solution. Once the first initial backup is complete the system then only backs up the deltas. Seeing as this is only doing a block level differential backup in essence you are getting a full backup each run but the data that is being stored on the Avamar device is only the changed blocks. This solution we have estimated will shrink our backups from 5 hours to 1 hour on our indexes. I'm sure it will provide similar results on vaults

EV INDEX Replication

So i infer from your response that a "restore" in this context would be to replicate the remote copy (if needed) to the lost site once the infrastrucuture at the lost site was back online?

Restore the relevant databases from backup.
Restore the servers etc and failback the cluster to the original primary active node?

Could we apply the same protection mechanism for databases - i.e. we are using Microsoft SQL and SQL mirroring and therefore avoid backups of the SQL databases as well?

Yes, replicate the remote

Yes, replicate the remote copy.

You could do the same with SQL, however, you should at least apply some sort of snapshot mechanism to protect against data corruption and stuff.

Also, you would have to break the mirror for upgrading EV:

There might be additional steps necessary to promote the secondary mirror as primary:

You might want to contact symantec support to inquire about other potential "issues" with SQL Mirroring.
(I haven't used it myself yet)


Quadrotech - EV Migration and Office 365 Solutions, creators of Archive Shuttle.