Joe_Budden
asked on
Database repair questions
Running Exchange 2007 SP2 CCR. Node1 is active, Node2 is passive. CMS is CCR1
Test situation-
We need to repair a database, but don't want to cause stress to the node the CMS is on.
1. Does running eseutil/p, eseutil/d, isinteg cause load on the server?
2. If we don't have space on the Node1, I assume we can copy the EDB file to the passive (node2) and run the eseutil utilities there?
3. In Isinteg, we need to specify a servername, is this the CMS name of the physical server name?
4. Based on the answer above, does this mean we can't run the Isinteg on the passive node?
5. Why does Isinteg need the CMS name?
Test situation-
We need to repair a database, but don't want to cause stress to the node the CMS is on.
1. Does running eseutil/p, eseutil/d, isinteg cause load on the server?
2. If we don't have space on the Node1, I assume we can copy the EDB file to the passive (node2) and run the eseutil utilities there?
3. In Isinteg, we need to specify a servername, is this the CMS name of the physical server name?
4. Based on the answer above, does this mean we can't run the Isinteg on the passive node?
5. Why does Isinteg need the CMS name?
ASKER CERTIFIED SOLUTION
membership
This solution is only available to members.
To access this solution, you must be a member of Experts Exchange.
ASKER
Thanks Meganuk.
Thus, if running ESEUtil/ISinteg hammers the disk, we should carry this out on the passive node, correct?
Do you know if there is any issue with running Isinteg on the passive node as long as the "-s servername" refers to the CMS name? Would that cause an issue though since the CMS is not on the passive? I didn't know you could enter the DB patch and not the -S for servername.
Thus, if running ESEUtil/ISinteg hammers the disk, we should carry this out on the passive node, correct?
Do you know if there is any issue with running Isinteg on the passive node as long as the "-s servername" refers to the CMS name? Would that cause an issue though since the CMS is not on the passive? I didn't know you could enter the DB patch and not the -S for servername.
Why do you need to repair the database anyway? Has the corruption replicated to the passive node? Do you have a DB down/dismounted?
Yes, you can run eseutil on the passive node, but you need to be prepared for copying the DB over the network and possibly doing that out of hours.
Isinteg used to work like that in the past so would be surprised if it doesn't allow that anymore, but then again MS have done some strange things over the years. I'll be able to test it tomorrow.
Yes, you can run eseutil on the passive node, but you need to be prepared for copying the DB over the network and possibly doing that out of hours.
Isinteg used to work like that in the past so would be surprised if it doesn't allow that anymore, but then again MS have done some strange things over the years. I'll be able to test it tomorrow.
Hmmm, isinteg requires the Information Store service to be running and in my experience that service is stopped on a passive CCR node
I tried isinteg without the -s <servername> an it won't let you run it without it. So, this goes back to my earlier question - why are you repairing the DB?
ASKER
Hi
This is just a hypothetical question in case the need comes up :) Just to better my understanding of ESEutil/ISinteg etc.
To confirm - the best method for fixing a DB on a cluster would be:
1. Copy the edb file to the passive node
2. Run ESEUtil /p and then ESEUtil/d
3 Copy the repaired edb file to the active node
4. Run ISinteg -fix
Reasons for running ESEUtil on the passive node is that it causes high resource usage. Reasons for running Isinteg on the active is because it needs the Information Store to be running.
Am I right?
This is just a hypothetical question in case the need comes up :) Just to better my understanding of ESEutil/ISinteg etc.
To confirm - the best method for fixing a DB on a cluster would be:
1. Copy the edb file to the passive node
2. Run ESEUtil /p and then ESEUtil/d
3 Copy the repaired edb file to the active node
4. Run ISinteg -fix
Reasons for running ESEUtil on the passive node is that it causes high resource usage. Reasons for running Isinteg on the active is because it needs the Information Store to be running.
Am I right?
Well in CCR you would only hopefully have 1 corrupt DB and the corruption hopefully won't have replicated over to the passive node, so you should be able to fail over and use the passive copy. Then Reseed back to overwrite the corrupt copy.
You need to do some testing on your hardware to determine how much eseutil and isinteg affect your setup because copying a DB will have network impact and may cause the server's clients to lose access or have slow access.
You need to do some testing on your hardware to determine how much eseutil and isinteg affect your setup because copying a DB will have network impact and may cause the server's clients to lose access or have slow access.
http://blogs.msdn.com/b/douggowans/archive/2009/03/04/offline-defrag-a-replica-database.aspx
If still you want to perform it. Than you need to seed the db again to passive node.
http://technet.microsoft.com/en-us/library/bb124706(EXCHG.80).aspx
Better to create new DB and move the mailboxes, then delete the old db. No risk, nno downtime.