Restoring Exchange in a Virtual Environment
If you can help it, just don't do it.
It's been a long time since I've had to restore an Exchange server at all, so it took me a while just to get the process down. Still, even after restoring Exchange 'by the book' I am running into errors that I can find almost no reference to anywhere on the web.
I believe it has something to do with the fact that the server was originally a physical machine which I converted to a VM and am now running in VMWare Workstation. The old volumes that previously contained the Exchange store are no longer available, so after cloning the Windows drive to the virtual machine I added the disks I needed and started the restore process.
Everything looks like it should be fine, but for some reason when I try to mount the Information Store Exchange tells me that the Hard Recovery option was not set on the restored database, even though I know for a fact that it was.
I thought that this was okay since I could still run the "eseutil /cc" command on it to manually set that option, but this fails with an error 0xC8000713. The only reference I found to this was here. This seems to be telling me that it can't load the restore.env file, although it could also be talking about the .mdb file itself... there isn't really any way to know yet.
In any case, I've had problems with either Yosemite Backup and/or Eseutil in regards to accessing the new volumes left and right. I'm guessing that it has something to do that the system is now running from a VMWare Virtual SCSI interface, instead of the old QLogic Fibre controller that it was using before. I would try the restore on the original system, but I have no way to add any external drives to it.
On top of all of this, when I boot the VM, the wmiprvse.exe service will randomly poll the c: drive to death for long periods of time which makes the machine virtually unusable. This also started after I moved the machine into the virtual environment and I have yet to solve the problem. That being said, I think I can at least alleviate it a little by moving the VM to a host machine that has a dedicated SCSI RAID volume availble for it instead of a SATAu RAID that is being shared with other things.
If I eventually work through this, I'll post my findings for the world (that is, all 2 or 3 of my fans) to see...
UPDATE
So I finally got around the WMI problem I was having by just Ghosting the original server to another identical machine. I just had to take two 146GB SCSI drives and stripe them so I would get enough space to run my restore of the Exchange database, as those machines have no way to attach anything other than fibre-based storage, externally.
After running the restore again, I was still unable to mount the database. I received the same error (at least one of them, anyway) in the Event Log as I did last time- Event 619. This basically says that Hard Recovery was not completed on the DB. After looking at the details of what ESEUtil was doing, I figured out that there were Exchange Log files missing that should have been restored.
For now, I've basically accepted the fact that YB doesn't seem to have grabbed all of my log files at the time it ran it's incremental backup of Exchange. I wasn't the person who was directly maintaining the backups at the time it was run, but in the past I had always preferred to have only full backups of Exchange for reasons just like this. In the time between the full backup and the incremental backup, I had moved a ton of mailboxes to this server and so there was about an additional 30GB added onto the database. This means about 5,000 additional log files are there, which may explain why not all of them were grabbed... it's a bit of an unusual scenario.
It's been a long time since I've had to restore an Exchange server at all, so it took me a while just to get the process down. Still, even after restoring Exchange 'by the book' I am running into errors that I can find almost no reference to anywhere on the web.
I believe it has something to do with the fact that the server was originally a physical machine which I converted to a VM and am now running in VMWare Workstation. The old volumes that previously contained the Exchange store are no longer available, so after cloning the Windows drive to the virtual machine I added the disks I needed and started the restore process.
Everything looks like it should be fine, but for some reason when I try to mount the Information Store Exchange tells me that the Hard Recovery option was not set on the restored database, even though I know for a fact that it was.
I thought that this was okay since I could still run the "eseutil /cc" command on it to manually set that option, but this fails with an error 0xC8000713. The only reference I found to this was here. This seems to be telling me that it can't load the restore.env file, although it could also be talking about the .mdb file itself... there isn't really any way to know yet.
In any case, I've had problems with either Yosemite Backup and/or Eseutil in regards to accessing the new volumes left and right. I'm guessing that it has something to do that the system is now running from a VMWare Virtual SCSI interface, instead of the old QLogic Fibre controller that it was using before. I would try the restore on the original system, but I have no way to add any external drives to it.
On top of all of this, when I boot the VM, the wmiprvse.exe service will randomly poll the c: drive to death for long periods of time which makes the machine virtually unusable. This also started after I moved the machine into the virtual environment and I have yet to solve the problem. That being said, I think I can at least alleviate it a little by moving the VM to a host machine that has a dedicated SCSI RAID volume availble for it instead of a SATAu RAID that is being shared with other things.
If I eventually work through this, I'll post my findings for the world (that is, all 2 or 3 of my fans) to see...
UPDATE
So I finally got around the WMI problem I was having by just Ghosting the original server to another identical machine. I just had to take two 146GB SCSI drives and stripe them so I would get enough space to run my restore of the Exchange database, as those machines have no way to attach anything other than fibre-based storage, externally.
After running the restore again, I was still unable to mount the database. I received the same error (at least one of them, anyway) in the Event Log as I did last time- Event 619. This basically says that Hard Recovery was not completed on the DB. After looking at the details of what ESEUtil was doing, I figured out that there were Exchange Log files missing that should have been restored.
For now, I've basically accepted the fact that YB doesn't seem to have grabbed all of my log files at the time it ran it's incremental backup of Exchange. I wasn't the person who was directly maintaining the backups at the time it was run, but in the past I had always preferred to have only full backups of Exchange for reasons just like this. In the time between the full backup and the incremental backup, I had moved a ton of mailboxes to this server and so there was about an additional 30GB added onto the database. This means about 5,000 additional log files are there, which may explain why not all of them were grabbed... it's a bit of an unusual scenario.
Comments
Feel free to visit my web-site: Heat And Glo Fireplace Inserts (Taipend.Com)