[prev in list] [next in list] [prev in thread] [next in thread] 

List:       toasters
Subject:    RE: exchange mailbox level recovery using NetApp
From:       "Rod Newcomb" <rodn () dice ! com>
Date:       2006-01-13 22:37:12
Message-ID: 93F92A00BB4D9F4ABFD302474A8AF0B80255E97F () sdmexchtest ! dice ! ad
[Download RAW message or body]

We've been working with Netapp and iSCSI for about a year now.  For the
most part, when their software works, it works very well.  But when it
doesn't it's pretty bad.

For example, we have a Clustered set of 920c's.   We use Qlogic HBA's
and GigE copper.  We've had multiple issues of when the MS cluster fails
over, we end up with multiple -1018 errors in our databases that need to
be cleaned up.   We've had three teams of consultants in to try to
determine what is causing the problem, but to no avail.

We're running 7.0.0.1 on the Filer (not a great release.. Were working
on upgrading).
Windows 2003 Enterprise SP1.
Exchange Enterprise SP1.
iSCSI initiator 1.06
QLOGIC 4010C HBA's.  2.0.0.5

We have a dedicated VLAN for our storage network, and no errors are
showing on the HBA's.    

The last issue we had, we couldn't recover our snapshot (verified with
eseutil as being good at the time of the backup) as it said the database
was still corrupted after recovering.   We ended up loosing about eight
hours of email for 30 users.

Also, beware that iSCSI is VERY expensive on space.  To get guaranteed
snapshots, you have to reserver 100% extra space of your lun.   That's
about 30% more than your average NFS file systems.    We didn't
understand that going into it, and grossly underestimated our storage
space needs.

We've had better luck with SQL, but still some issues.

In the lab, Single Mailbox restore worked very well.  It basically
"virtually" mounts the message store and you select the information you
want to bring back into your production database.  Exchange isn't the
wiser.  

I just wish we could get down to the bottom of our database issues
involving iSCSI.  The speed is incredible, and iSCSI is a great solution
for our company on paper.  In practice, it's been a little different
experience.

Rod Newcomb



-----Original Message-----
From: owner-toasters@mathworks.com [mailto:owner-toasters@mathworks.com]
On Behalf Of Leeds, Daniel
Sent: Friday, January 13, 2006 4:02 PM
To: Brian Parent; Netapp List
Subject: Re: exchange mailbox level recovery using NetApp




On 1/13/06 1:19 PM, "Brian Parent" <bparent@calvin.ucsd.edu> wrote:

> We're considering using NetApp storage via iSCSI LUNs for Exchange 
> email databases, but are concerned that we'll run into trouble due to 
> some obscure/hidden Microsoft magic.
> 
> I suspect that the way NetApp does snapshots, and presents the disk as

> a LUN to the exchange server, it prevents any such problems.  That is,

> there is no way for the M$ software to know the difference between the

> live production data, and the snapshot data from which you're making 
> mailbox level recovery.
> 
> Anyone actually had success with such a setup?
> 

Netapp makes a product called SnapManager for Exchange.  Look into this
product.  It allows things such as single mailbox restores and more.


[prev in list] [next in list] [prev in thread] [next in thread]