What value does having SnapManager installed on the Exchange server add?
By Hannah Rich on Jan 14, 2011 at 11:09 AM in Engineer Knowledgebase, Customer Education, Did You Know Series

First: here’s what SnapManager for Exchange (SME) provides:
-It backs up the database through Exchange, meaning your transaction logs are rolled in
-It quiesces the database, ensuring that the NetApp snapshot is clean. If this doesn’t happen, you get a dirty backup.
-It allows you to schedule NetApp snapshots, automatically delete them once they’ve expired, and validate backups
-It’s also great for restoring your database, and stores information that lets you do Single Mailbox Recovery.
The key takeaway is that snapshots are the underlying backup technology, SME is the tool to configure and schedule those snapshots, coordinate backups and restorations, provide notifications and alerts, and to allow for instant recovery. It’s how you actually get value from the snapshots.
I would definitely skip hourly snapshots, unless you really require that kind of granularity. For example, at most clients, we do 6AM / 6PM / Midnight snapshots. I try to avoid noon snapshots, as Outlook clients register any lag over 5 seconds as “disconnected”, and this can scare users. Your databases don’t go offline, but Outlook is so intensely sensitive that it freaks people out.
With the 3 snapshots a day schedule recommended above, you can go for over 2 months of snapshots without deleting any. You want to DISABLE any scheduled snapshots you may be taking, as they are worthless if the database hasn’t been quiesced (IO frozen). SME should be the one initiating the snapshots, as that ensures you have a clean database and ensures that the Exchange backup procedure rolls the logs into the database.
For you, the key is going to be Single Mailbox Recovery. While you have your tape backups, those can take forever to restore from. You can literally have any snapshot of your Exchange DB/logs mounted in under 5 minutes once you know what you’re doing. You can then poke around in the datastore with the SMBR tool from NetApp, and you can pull individual emails or mailboxes out to restore from. Doing that from tape would literally be a nightmare, if not impossible.
To do:
1) Disable all scheduled snapshots on both DB and logs volumes
2) Run through the Backup Wizard of SME, the very last step will say “Schedule”, schedule it there (uses Windows Task Scheduler, use the “Multiple Schedules” option to keep it in one job)
3) Configure ‘Notification Settings’ in SME to email the proper CIO and CoSM email addresses on failures (check Advanced settings for “On failure only”), you’ll need an open SMTP relay
4) Make sure you have plenty of open volume space. I suggest 3-to-1 or 2-to-1 with your LUN, so 100GB volume for a 30GB LUN, or 100GB volume for a 50GB LUN, you get the idea. Monitor as the snapshots roll in to see if you are overgrowing your volume.
Chase Christian – CIO Engineer