Hi All

Ive been thinking about trying to drop the mailbox backups for our new Exchange 2010 environment as they are always so slow and hit the server.  So ive been thinking about just doing database backups and using the offline mining tool to be able to pull out individual mailboxes from old copies of the database :-

http://documentation.commvault.com/commvault/release_8_0_0/books_online_1/english_us/features/mining_tool/ex_mining_tool/ex_mining_tool.htm

Has anyone done this before?  Ive read the doco a couple of times and im not sure if you can run the mining tool on any system ie Member server, or does it need to be on an exchange server?  can it just read the core files (.edb for example)?

Also not sure if offline mining works with 2010?

Im hoping that soon we will be able to do database backups and have CV collect the metadata within that backup so you can look into it and get out the individual files you want, like the way the VADP backups work with 'Collect Metatadata' option on.

bugger.  So im hoping that in V9 or 10 they can drop Mailbox backups.....

It is possible that we will support this in the current release but I do not have any information regarding this.  I will update this thread if this happens.

great, thanks Patrick :)

Hi there,

We were also thinking of dropping the mailbox backups because of the speed and the hit on the exchange server... Did some testing and turned to incrementals instead...

On full on weekends, incremental during the week... works fine!

Can I ask what is your average throughput on the Mailbox backup?

I am getting around 3GB/Hr; while Exchange DB backup gets over 20GB/Hr.

Hi Rtsung,

Exchange Mailbox Backups average around 2-4 gigs an hour. The limitation here is on the Messaging APIs that are used for the backups. The ammount of API calls required to backup at this granular level is where the bottle neck for this backup type is.

One recommendation is to create multiple subclients for the Exchange Mailbox backups. Each subclient makes its own connection to the Exchange Server and has the potential to run at about 2-4 gigs per subclient when ran simultaneously. If you go this route, I would usually recommend 3-5 subclients to start as each subclient will be launching a seperate backup process on the server with the Exchange Mailbox iDA installed.

Thanks!

Just to update this thread regarding the mining tool.

The exchange mining tool now supports exchange 2010 in commvault 9.0.

thanks patrick, we are on CV V9 now too.

So does my initial post and questions apply still as a correct method?