service specific error 1011

[mbrenowitz]

My company uses a cobbled together box for their Exchange 5.5 server. It has one hard drive, split into two partitions, C: (1GB, FAT) and D: (7GB, NTFS). Last night the hard drive shut itself off, after several SCSI controller errors; the server room has also been experiencing heat problems. It is running ok now, but on reboot, the Information Store service starts at boot, but immediately shuts down. Event Viewer shows service specific error 1011, event code 7024. I haven't been able to find this error listed in any Microsoft help documentation, on their web site, any other tech support site, or in Exchange books. The server is running Windows NT 4.0, service pack 6a. Because the Information Store service won't start, neither will the Event Service or the Internet Mail Service. The message stores, folders Dsadata, Mdbdata, and Mtadata sit on the D: drive. I shut off all Exchange services, then copied all Exchange folders and files to our file server as an emergency backup.
Any suggestions would be greatly appreciated.
Marc

 

[Zelandakh]

Marc,

If an Exchange server goes down without being shut down properly, the GUIDs are synchronised so when you restart you are missing what is basically the transaction log (if you know SQL server). To repair this error, do the following:

Ensure that the Directory and System Attendant Services are running. If either of these services are not running, Isinteg will fail with a DS_COMMUNICATIONS_ERROR message.

Change to the d:\Exchsrvr\Bin directory, and type Isinteg -patch. The GUIDs will be replaced, and Isinteg will report that the Information Store has been updated successfully.

Restart the Information Store service. Hopefully this will cre your problem.

 

[Wibble]

i'd be very careful before running the patch. If a machine is not shut down properly it may not necessarily be an issue of GUID's needing to be brought back inline. This usually happens when you restore an OFFLINE backup.

Have you checked the transaction logs???? is it failing to create them?? is it failing on a particular transaction log update??? You may want to check the checkpoint file to see where it thinks it has commited up-to...An abnormal shutdown will cause the transaction logs to be checked and rolled forward when restating the service....you may want to find the last good log, remove any logs after it, delete the edb.chk rename the last good log to edb.log and start the service up again.......I haven't done this in ages so check Q240145.

hope this helps

 

[Zelandakh]

Good point. It is usually an off line issue, but I have seen it on abnormally shut down servers.

EDBUTIL can help using the /pub switch I think.

Try Wibble's suggestion first and the worst that will do is nothing, but rename files instead of deleting them.