BUG: Improper Load of Master Hangs SQL ServerID: Q119912
|
If you load an outdated dump of the master database, it may cause
Microsoft SQL Server to not restart.
Bug number 866 covers a specific case of the problems that can occur if an
incorrect dump of master is loaded. If you drop a database and
recreate it so that it gets the same database id, and the log of the new
database maps back into the same area on the same device that held the old
database, then SQL Server will not restart if a dump of master is loaded
that contains the old mapping for the database.
SQL Server version 4.2 generates a Dr. Watson window on the restart
which reads:
SQLSERVR.EXE had a stack overflow (0xc00000fd), address:0x0041f27f.
If you have a copy of the MASTER.DAT from just before the load, restore
it and restart SQL Server. If you have another dump of master that is
good, restart SQL Server from the command line as follows:
sqlservr -T3608
Microsoft has confirmed this to be a problem in Microsoft SQL Server
version 4.2. We are researching this problem and will post new information
here in the Microsoft Knowledge Base as it becomes available.
This problem does not occur in Microsoft SQL Server version 6.0.
As mentioned in the SQL Server "System Administrator's Guide," you should always dump your master database whenever any of the system information in the database changes, and any attempt to load the master database must be made from this most current dump. Page 309 of the "System Administrator's Guide" lists the most common commands which alter system information in the master database.
Additional query words: Windows NT
Keywords : kbother kbbug4.20 SSrvWinNT
Version : 4.2
Platform : WINDOWS
Issue type :
Last Reviewed: March 20, 1999