Neither - using NAS.

-----Original Message----- From: McCarthy, Tim [mailto:timothy.mccarthy@netapp.com] Sent: Friday, April 01, 2005 6:46 AM To: Holland, William L Subject: RE: Exchange 5.5 and OnTap 7.0.0.1 Problem?

You did not mention in your post if you were using VLD or iSCSI, your version of SnapManager for Exchange and your version of SnapDrive.

Have you checked the compatibility Matrix on NOW?

--tmac ======================

-----Original Message----- From: Holland, William L [mailto:hollandwl@state.gov] Sent: Friday, April 01, 2005 5:59 AM To: 'toasters@mathworks.com' Subject: Exchange 5.5 and OnTap 7.0.0.1 Problem?

The F825 hosting the message stores were upgraded to OnTap 7.0 on 3/5/2005 and to 7.0.0.1 on 3/23/2005. Also, on 3/23/2005 we started SnapMirror'ing the message stores to an R200 running 7.0.0.1. On the afternoon of 3/28/2005 we saw one server cycle its Exchange services a few times and on 3/29/2005 both Exchange servers were cycling their Exchange services. By 3/30/2005 it had become so bad and so frequent we made the decision to move the message stores to the local storage to isolate if the problem was with the filer or with the databases themselves. All is running fine now.

On 3/28/2005 I conducted a test to determine if backups performed with SnapManager (snapbexc) were truly recoverable. To do so I built a new Exchange 5.5 SP4 server (same organization and site name, but did not join existing site and did not configure any connectors) on NT4 SP6a, connected it via direct GbE to an F810 running OnTap 7.0.0.1 and used ndmpcopy to copy from the snapshots on the F825 to the F810. When I tried to start the Directory Store I received an error message. I used ESEUTIL to repar the database and when I tried to start it then I was informed that it would not start as the computer I was attempting to start it on was not the same name as the one it had been created on. At this point I shut this server down and deleted the files and volumes created for the test. It was within minutes of this that we saw the first cycle of the Exchange services on the server whose database I was using a copy of to test. I could almost believe that this could have caused a problem on one server but not two and then also given the fact that now that we have moved the message stores to local storaget they are running fine. But, in any event, I've learned not to discount anything as possible in a networked environment.