Video Screencast Help

EV 10 and Exchange 2010

Created: 01 Jan 2013 • Updated: 01 Jan 2013 | 8 comments
This issue has been solved. See solution.

 

 
We are planning to have 1 EV site with 5 EV servers, 1 SQL server for targetting 5 exchange servers (2010). I would like to know that with the concept of DAG in exchange 2010 in which the databses can failover to any mailbox servers and after the provisioning task has run archivig will be processed by the Mailbox Archiving task associated with the new Exchange Mailbox Server.
 
My dout is if we have san storage connected to each EV server which will have partitions created on it respectively.
eg.
 
EV A having SAN storage A connected will archive  mailboxes of Mailbox server A (Database A,Database B,Database C )  
EV B having SAN storage B connected will archive  mailboxes of Mailboxes B (copy of Database A,copy of Database B,Database D )  
 
If the Database A and B which were active on Mailbox server A failovers over to Mailbox server B and provisioning task is run after the failover and EV B starts archiving  Database A and Database B which are now active on Mailbox Server B, then in this case the archived items will be stored in storage A or Storage B and how ?

Comments 8 CommentsJump to latest comment

JesusWept3's picture

If the users are on Exchange A targeted by EVServerA and then fails over to ExchangeB where the task resides on EVServerB , when items are archived it will still write to the users vault store

So an item that is archived will profess through the MSMQ's like

EVserverB\Exchange ServerB A2
then to EVServerA\Storage Archive Queue
then to EVServerB\Exchange ServerB A1

So no matter where the exchange archiving task resides, or what Exchange Server the user is on, there items will ALWAYS be written to the Vault Store they belong to when they were first enabled

JesusWept3's picture

Oh sorry, didn't see saying Copies. then yeah only Active Databases will be archived, not the copies etc
But anyho, based on that guide, if a DAG failover occurs and the users mailbox is now being archived by a task on another server, the following applies (which i mentioned originally)

"Yes, it is possible that performance would be affected. For example if you have a scenario where you have a 1:1 relationship of 1 Enterprise Vault server per Exchange Mailbox Server and with a DAG configuration similar to that of the customer scenario in this document, if MailboxDatabase1 fails over to MailboxServer2, once the provisioning task has run, the mailboxes in MailboxServer2 will now be processed by the second EV server as that is the server that hosts the archiving task for MailboxServer2. In addition to already archiving mailboxes in MailboxDatabase3 and MailboxDatabase4, the second EV server now has to process mailboxes in MailboxDatabase1 and MailboxDatabase2. In that event the Enterprise Vault server will have more mailboxes to process, and because the storage service associated with archives relating to mailboxes in MailboxDatabase1 and MailboxDatabase2 will reside on the first EV server, archived data will have to be transferred over MSMQ to the first EV server. "

Danish A's picture

I am new to EV so can you please explain me in more details "the storage service associated with archives relating to mailboxes in MailboxDatabase1 and MailboxDatabase2 will reside on the first EV server, archived data will have to be transferred over MSMQ to the first EV server. " 

Since we will have 1:1 relationship of 1 EV server per Mailbox server, how should we map the storage drives on each EV servers, since as per my understanding storage drives on EV A will not be available on EV B server, I need to understand the process by which MSMQ will transfer the data to first EV server when the mailboxes are processed on second EV server.

 

JesusWept3's picture

Enterprise Vault Partitions that belong to Vault Stores will be assigned per EV Server

So you would have something like

VaultStore1 (on EVServer1)
Partition1 -> E:\Partition1
Partition2 -> F:\Partition2

VaultStore2 (on EVServer2)
Partition1 -> E:\Partition1
Partition2 -> F:\Partition2

Let's say you have Exchange1's Archiving Task on EVServer1 and Exchange2's Archiving Task on EVServer2

You enable Archiving for a user on Exchange1, EV then creates an archive on VaultStore1

You enable Archiving for a user on Exchange2, EV then creates an archive on VaultStore2

If a user gets moved from Exchange1 to Exchange2, although the users Task that archives the users mailbox is on EVServer2 , the users archive still belongs to VaultStore1

When an item gets archived in the users mailbox, it gets passed from EVserver2 to EVserver1 to store the item in the archive

Danish A's picture

Dear,

Really appreciate your explanation, now most of my douts are clear except the process by which MSMQ will transfer the data from VaultStore2 on EVServer2 to fVaultStore1 on EVServer1 .

JesusWept3's picture

Ok so it does the following

EVServer2\Private Queues\Archiving Task for Exchange2 A2

EVServer2\Outgoging Queues\Storage Archive on EVServer1

EVServer1\Outgoing Queues\Archiving Task for Exchange2 on EVServer2

So the way archiving works is it goes to A2 queue first that puts the item in to Pending state and tells the task to get a copy of the message

It then goes from A2 to Storage Archive and takes the copied message and turns it in to an EV message (DVS files etc)

Once it's been processed it goes from Storage Archive to the A1 Queue and "post processes" the item meaning it will turn it in to a shortcut

So to recap

Exchange Task A2 (pending) -> Storage Archive -> Exchange Task A1 (post process)

When the task and vault store are on the same server it all uses the Private Queues

When the task and vault stores are on separate servers then EV uses the "Outgoing Queues" to connect to a queue that resides on another server

Danish A's picture

Dear ,

 

Thanks for your explanations it was really great and cleared all my douts.

 

Regards,

 

Danish