Exchange Database Corruption

Dmitry

unread,

Sep 16, 2005, 1:01:02 PM9/16/05

to

I am running the Exchange Server 2000 SP3
Exchange database backup returned the following error (I am using the CA
Brightstor Backup 11.1)
"Exchange Server Error. HrESEBackupEnd failed."
Brick level backup (CA Brightstor Backup 11.1) works well. No errors are
generated.
NT Exchange backup returned the following error:
WARNING: Portions of “\Mailbox Store (MAIL)” cannot be read. The backed up
data is corrupt or incomplete

During database backup, (with ether NT Exchange or CA backup software)
application log fills with the following events:

Event Type: Error
Event Source: ESE
Event Category: Logging/Recovery
Event ID: 217
Date: 14/09/2005
Time: 6:48:56 AM
User: N/A
Computer: MAIL
Description:
Information Store (2692) bb119f70-0b58-4a1f-8aab-6634ad37f630: Error (-613)
during backup of a database (file C:\Program
Files\Exchsrvr\mdbdata\priv1.stm). The database will be unable to restore.

Event Type: Error
Event Source: ESE
Event Category: Logging/Recovery
Event ID: 478
Date: 14/09/2005
Time: 6:48:56 AM
User: N/A
Computer: MAIL
Description:
Information Store (2692) The streaming page read from the file "C:\Program
Files\Exchsrvr\mdbdata\priv1.stm" at offset 1245716480 (0x000000004a402000)
for 4096 (0x00001000) bytes failed verification due to a page checksum
mismatch. The expected checksum was 2777780870 (0xa5919286) and the actual
checksum was 2814150562 (0xa7bc87a2). The read operation will fail with
error -613 (0xfffffd9b). If this condition persists then please restore the
database from a previous backup.

Event Type: Error
Event Source: ESE
Event Category: Logging/Recovery
Event ID: 215
Date: 14/09/2005
Time: 6:48:56 AM
User: N/A
Computer: MAIL
Description:
Information Store (2692) bb119f70-0b58-4a1f-8aab-6634ad37f630: The backup
has been stopped because it was halted by the client or the connection with
the client failed.

Event Type: Information
Event Source: ESE
Event Category: Logging/Recovery
Event ID: 222
Date: 13/09/2005
Time: 4:24:08 PM
User: N/A
Computer: MAIL
Description:
Information Store (2628) bb119f70-0b58-4a1f-8aab-6634ad37f630: Ending the
backup of the file C:\Program Files\Exchsrvr\mdbdata\priv1.stm. Not all data
in the file has been read (read 299630592 bytes out of 1855987712 bytes).

Event Type: Error
Event Source: NTBackup
Event Category: None
Event ID: 8012
Date: 13/09/2005
Time: 4:24:08 PM
User: N/A
Computer: MAIL
Description:
The 'Microsoft Information Store' returned 'Error returned from an ESE
function call (d).
' from a call to 'HrESEBackupRead()' additional data '-'

Event Type: Error
Event Source: NTBackup
Event Category: None
Event ID: 8012
Date: 13/09/2005
Time: 4:24:08 PM
User: N/A
Computer: MAIL
Description:
The 'Microsoft Information Store' returned 'Error returned from an ESE
function call (d).
' from a call to 'HrESEBackupClose()' additional data '-'

So, I have all the indications of the database corruption. I ran the
checkdisk on the hard drive. Checkdisk returned errors, but all errors were
successfully corrected. Just to be sure, I have used disk imaging software to
“Clone” the troubled hard drive into a good one. Exchange server started
without any problems, but database still had errors. The last good database
backup is a week old now. I do have a recent brick level backup.
Next, I made an attempt to repair the Exchange database using eseutil, and
isinteg utilities with the following switches:
Step 1: eseutil /P “c:\Program Files\exchsrvr\mdbdata\priv1.edb”
/tc:\tempdb.edb
Step 2: eseutil /D “c:\Program Files\exchsrvr\mdbdata\priv1.edb”
/tc:\tempdb.edb
Step 3: isinteg –s mail –fix –test alltests

Eseutil found some errors in the database and presumably fixed it. But the
.stm file still had some corruption, and backup failed again. Reading the
Exchange blog:
http://blogs.technet.com/exchange/archive/2004/06/18/159413.aspx
I learned, that .stm file could be re-created. Since all of our users use
Outlook (MAPI protocol) I made an assumption, that my .stm file does not
store any useful data. So, I ran the eseutil again:
eseutil /P “c:\Program Files\exchsrvr\mdbdata\priv1.edb” /tc:\tempdb.edb
/createstm
Big mistake. It wiped out everything, what came into the system using SMTP
protocol. Luckily, before running the eseutil, I made another “Clone” of the
hard drive. I reverted to the good clone of the hard drive, which saved the
day.

So, I have the following questions:

- Do I need to continue recovery attempts using eseutil and isintegr, or
just restore the good copy of the database, and apply all the recent changes
using the brick level backup?
- Does the eseutil utility repair the .stm file?

Thanks for all comments and suggestions

Dmitry

Susan

unread,

Sep 16, 2005, 1:08:01 PM9/16/05

to

are you running Enterprise Edition? if so, I would create a new mailbox
store and move the users to it...then try backing up the new mailbox
store...this does appear that you've got corruption, and you might need to
run eseutil and isinteg against the store, but if at all possible, I'd move
the mailboxes, first...or contact MS PSS on this...

"Dmitry" <Dmi...@discussions.microsoft.com> wrote in message
news:DB54BBAC-7078-48FC...@microsoft.com...

Andy David - MVP

unread,

Sep 16, 2005, 1:11:49 PM9/16/05

to

Recovery from a good backup should always be your first choice.
eseutil shold only be used to repair when you have no other option.
THe repair process is not always successful and you may lose data.

the stm can be repaired with eseutil in theory.