Link to home
Start Free TrialLog in
Avatar of ilguybob
ilguybobFlag for United States of America

asked on

cannot mount information store after restore of database

I am doing some disaster revovery planning for our exchange server. I captured a vmware image of the exchange server, and set it up as a virtual machine (disconnected from our LAN). The vm exchange server came up fine and all exchange services will start, however, the store will not mount. I figured that in a disaster recovery scenario, we would have to restore the most recent tape backup of the exchange information store anyway, I took the latest such backup and restored it to the vm exchange server. Again, the store would not mount. I am getting the following error in the event log:
Event Id 457
Source: ESE
Category: logging/recovery
"The log signature of the existing logfile E00.log doesn't match the logfile from the backup set. Logfile replay cannot succeed unless all signatures match."

I read Microsoft article KB823016, which states that this can happen if the Exchange information store service is not running when the restore was performed. I was pretty sure it was running, but to be sure I once again took a backup of the "real" exchange server and restored it to the vm exchange server, this time making sure the service was running. Afterwards, I had the same result and the same event errors.
I also have alot of other errors in the event logs:

Event ID: 904
Source: ESE Backup
"callback fnction call ErrESECBRestoreComplete ended with error 0xC8000262 existing logfile has bad signature."

Event ID: 9175
Source: MSExchangeSA
"The MAPI call 'openmsgstore' failed with the following error. The Microsoft Exchange server computer is unavailable"

Along with a whole slew of other errors relating to exchange.

NOTE:
I also tried running exchange setup with the /domainprep switch.
Also reapplied Exchange SP2 and Windows SP2.

OUR ENVIRONMENT:
Exchange server 2003 SP2 on Windows 2003 SP2
Two domain controllers, which are also running as vm servers on the same box as the vm exchange.
There appears to be good communication between the vm servers on that box.
Any ideas?
ASKER CERTIFIED SOLUTION
Avatar of rakeshmiglani
rakeshmiglani
Flag of India image

Link to home
membership
This solution is only available to members.
To access this solution, you must be a member of Experts Exchange.
Start Free Trial
Avatar of ilguybob

ASKER

State is "dirty shutdown"
I have a restore.env file at C:\temp\first storage group

I should have mentioned, the public store is still in the default location. The private store is located at D:\databases and the log files are at D:\mddata. This is the same as the production exchange server. I read somewhere about file permissions being a possible cause so I gave everyone full controll to those folders to see if it had any effect.
I also tried deleting all files out of those locations before doing a restore.
did you try running eseutil /cc against restore.env?

eseutil /cc C:\temp\first storage group
I ran that and got this:

log files range: E00101EC.log - E0010205.log
last restore time: Wed Dec 31 18:00:00 1969

Recover status: recoverended
recover error: 0xc8000262
recover time: Tue Jul 08 17:36:21 2008

Operation terminated with error -939523486 (existing logfile has bad signature.)
any errors in the application logs on this server?
Yes, the ones I stated in the original post, as well as a few others:

Event id: 929
Source: msexchange transport

event id: 619
Source: ESE


did you have any log files in the log files location prior to doing the restore?
can you completely clear out all the transaction logs and checkpoint file and also the databases from the respective folders on this server and do a restore again after selecting the "this database can be overwritten by  a restore" in the mailbox store properties?
I have done that before, but I guess it will not hurt to try again. We have a large database, about 45gig, so it will take several hours.
Do you know how much free space I need to do this? I have 1.7GB free on C and 36GB free on D (with the database files included). Does it need temp space?

And FYI, I cannot backup the store on the vmserver as it is now. It says "instance not found" in the backup log. That is one of the steps in the Microsoft article  KB823016 I listed above.
is this a test/recovery server?
if yes, then you can delete the databases that you had restored earlier as you already have them on a backup take which you will be restoring now.
why do you want to backup the data that you have on this vmware box?
Yes, it is a test server.
I attempted to back it up because that is one of the steps in the MS article that I referenced.
a backup of this test box is not required as we are doing a restore again and you already have the data in a backup tape
just make sure that there are no existing logs in the folders before you start the restore.
the database is around 45gb, so you need that much of free space + some more for the log files.
Avatar of Naga Bhanu Kiran Kota
Hi,

First things first. What is the backup software you are using to restore the data.

You have mentioned that the free space is "Do you know how much free space I need to do this? I have 1.7GB free on C and 36GB free on D (with the database files included). Does it need temp space?"

where as the size of your IS is 45 GB.

Is that after already restoring the files from earlier backup or is that after clearing out the database from exchange.

it certainly needs temp space as that of the size of your backup. So ensure that you have sufficient temp space. The backup solution will give you the option to change the temp space location from c:\temp to d:\temp or if you can add another virtual disk e:\

It seems to only restoring the edb and stm files not the log files.

bhanu
I think I may have forgotten a key piece of info here.

When I originally took the vmware image of the exchange server and configured it, I tried to boot and it presented me with the activate windows screen. Of course I could not log in or do anything, so my only option was to do a repair install of windows 2003 server from the cd. from what I have been reading, this is what srewed everything up. This is a Dell server with an oem license. Is there any way around this? I called Microsoft to reactivate and of course they would not give me another activation code, even though I told them it was just for DR testing.

If I get this one activated, will it "deactivate" the original, production exchange server?
Hi,

When you try to boot the vmware image VMware pops a message if you want to keep the uuid or create a new one if you keep it, it will not windows re-activation.

As far as your Activation is concerned you need to check the license agreement. As far as license agreement is concerned i think you can certainly get a new activation code if you are re-building the OS as a Virtual machine. As MS license agreement for win 2003 provides you license upto 4 VMs if you have a valid 2003 license. should be enterprise i believe.

so try to recreate a vmimage and when it asks you to create a new uuid let it keep the uuid. The uuid id basically hardware id. Windows need re-activation if there is a hardware change.

bhanu
I got a new vmware image and never did see any mention of uuid.
Anyway, I tried mounting the store this time and got the error saying that the database is in an inconsistent state.
After a repair of the database, everything was ok.
When restoring Exchange database, it is important that there are no logfiles present in the log folder.
When starting the MS Information Store service, which you would normally do before restoring , it will try to mount the databases, and some new logfiles are generated.
These MUST be deleted before running the actual restore process.
If not, you will get the error "0xC8000262 existing logfile has bad signature"
Took me two whole days to find out :-), but then my restore (From 3. party backup) was succesfull.