November 3, 2008 at 10:43 am
SQL2000 environment.
Starting the SQL services fails. It just hangs at the "starting..." phase, goes back to being stopped.
Here's the latest log file:
008-11-03 10:46:41.67 server Copyright (C) 1988-2002 Microsoft Corporation.
2008-11-03 10:46:41.67 server All rights reserved.
2008-11-03 10:46:41.67 server Server Process ID is 1492.
2008-11-03 10:46:41.67 server Logging SQL Server messages in file 'C:\Program Files\Microsoft SQL Server\MSSQL$BKUPEXEC\LOG\ERRORLOG'.
2008-11-03 10:46:41.71 server SQL Server is starting at priority class 'normal'(8 CPUs detected).
2008-11-03 10:46:41.96 server SQL Server configured for thread mode processing.
2008-11-03 10:46:41.96 server Using dynamic lock allocation. [500] Lock Blocks, [1000] Lock Owner Blocks.
2008-11-03 10:46:42.03 spid3 Starting up database 'master'.
2008-11-03 10:46:42.35 spid3 0 transactions rolled back in database 'master' (1).
2008-11-03 10:46:42.35 spid3 Recovery is checkpointing database 'master' (1)
2008-11-03 10:46:42.45 server Using 'SSNETLIB.DLL' version '8.0.766'.
2008-11-03 10:46:42.45 spid5 Starting up database 'model'.
2008-11-03 10:46:42.46 spid3 Server name is 'GWSQL2\BKUPEXEC'.
2008-11-03 10:46:42.46 spid3 Skipping startup of clean database id 4
2008-11-03 10:46:42.46 spid3 Starting up database 'BEDB'.
2008-11-03 10:46:42.64 spid3 Error: 9003, Severity: 20, State: 1
2008-11-03 10:46:42.64 spid3 The LSN (1382:341:1) passed to log scan in database 'BEDB' is invalid..
2008-11-03 10:46:42.68 spid3 Error: 3414, Severity: 21, State: 1
2008-11-03 10:46:42.68 spid3 Database 'BEDB' (database ID 5) could not recover. Contact Technical Support..
2008-11-03 10:46:42.70 spid5 Clearing tempdb database.
2008-11-03 10:46:42.70 server SQL server listening on 192.168.0.20: 1137.
2008-11-03 10:46:42.70 server SQL server listening on 127.0.0.1: 1137.
2008-11-03 10:46:42.71 server SQL server listening on TCP, Shared Memory, Named Pipes.
2008-11-03 10:46:42.78 spid5 Starting up database 'tempdb'.
2008-11-03 10:46:42.81 spid3 Recovery complete.
2008-11-03 10:46:42.81 spid3 SQL global counter collection task is created.
2008-11-03 10:46:44.98 server SQL Server is ready for client connections
2008-11-03 10:46:48.62 spid51 Error: 947, Severity: 16, State: 1
2008-11-03 10:46:48.62 spid51 Error while closing database 'BEDB' cleanly..
2008-11-03 10:46:48.79 spid51 Starting up database 'BEDB'.
2008-11-03 10:46:48.87 spid51 Error: 9003, Severity: 20, State: 1
2008-11-03 10:46:48.87 spid51 The LSN (1382:341:1) passed to log scan in database 'BEDB' is invalid..
2008-11-03 11:36:39.61 spid51 Using 'xpstar.dll' version '2000.80.760' to execute extended stored procedure 'sp_MSgetversion'.
2008-11-03 11:36:39.70 spid51 Starting up database 'msdb'.
Any guesses?
November 3, 2008 at 10:55 am
Just out of curiusity; was this database
The LSN (1382:341:1) passed to log scan in database 'BEDB' is invalid..
just recently being restored??
Cheers,
John Esraelo
November 3, 2008 at 11:30 am
Hi,
I think you may have more than one problem. Obviously your BEDB is in a bad shape but that shouldn't stop your service from coming up.
Can you start the SQL Server in minimal mode and see what happens? Go to the BINN directory of that instance and run:
sqlservr.exe -c -m -T3608
The 3608 traceflag skips recovery for all databases except master. If you are on a cluster you need to take the SQL Server group offline first.
/Elisabeth
elisabeth@sqlserverland.com
MCITP | MCT
http://sqlblog.com/blogs/elisabeth_redei/
http://linkedin.com/in/elisabethredei
November 3, 2008 at 11:52 am
Elisabeth Rédei (11/3/2008)
Hi,I think you may have more than one problem. Obviously your BEDB is in a bad shape but that shouldn't stop your service from coming up.
Can you start the SQL Server in minimal mode and see what happens? Go to the BINN directory of that instance and run:
sqlservr.exe -c -m -T3608
The 3608 traceflag skips recovery for all databases except master. If you are on a cluster you need to take the SQL Server group offline first.
/Elisabeth
That was my thinking.
When trying to start up the server in minimal mode I get:
03 12:43:14.01 server Attempting to initialize Distributed Transaction Coordinator.
2008-11-03 12:43:16.07 spid3 Warning ******************
2008-11-03 12:43:16.07 spid3 SQL Server started in single user mode. Updates allowed to system catalogs.
2008-11-03 12:43:16.07 spid3 Recovering only master database
2008-11-03 12:43:16.07 spid3 Starting up database 'master'.
2008-11-03 12:43:16.09 spid3 Error: 823, Severity: 24, State: 2.
2008-11-03 12:43:16.09 spid3 Error: 3313, Severity: 21, State: 2.
2008-11-03 12:43:16.11 spid3 Cannot recover the master database. Exiting.
I have their IT people trying to restore the master backup from last night right now......
I'm obviously not a DBA, but have bungled my way through enough things to be reasonably apt here and there. As a result I imagine that I am missing some fundamental knowledge, i.e. - will restoring the master database from an older backup hose everything up?
Best practice?
November 3, 2008 at 11:54 am
John Esraelo (11/3/2008)
Just out of curiusity; was this databaseThe LSN (1382:341:1) passed to log scan in database 'BEDB' is invalid..
just recently being restored??
This is a customer's (almost entirely unmanaged) infrastructure. I can't imagine that with his lack of SQL knowledge that he would be doing anything with ANY of the databases.
Other than that that's all I know about it.
November 3, 2008 at 12:02 pm
I see..
what's interesting about the error message is that it normally revolves about the (Error : 9003 severity 20, State 1) master database having issues with and not the DB in question.
Unfortunately, in most cases similar to this error message the DBA needs to restore the master database and any other changes required that may be lost between the last master db backup and the present master db restore.
That's is the reason why it has been recommended to backup the system databases before and after each and every sql server change.
Good luck!
JohnE
Cheers,
John Esraelo
November 3, 2008 at 12:07 pm
John Esraelo (11/3/2008)
I see..what's interesting about the error message is that it normally revolves about the (Error : 9003 severity 20, State 1) master database having issues with and not the DB in question.
Unfortunately, in most cases similar to this error message the DBA needs to restore the master database and any other changes required that may be lost between the last master db backup and the present master db restore.
That's is the reason why it has been recommended to backup the system databases before and after each and every sql server change.
Good luck!
JohnE
My question is what exactly does the master db contain in terms of changes....
November 3, 2008 at 12:12 pm
Pinal Dave is one of the guys...
follow the link..
Cheers,
John Esraelo
November 3, 2008 at 12:13 pm
It should only be your logins unless someone put user objects in there (which is not normally not the case).
/Elisabeth
elisabeth@sqlserverland.com
MCITP | MCT
http://sqlblog.com/blogs/elisabeth_redei/
http://linkedin.com/in/elisabethredei
November 3, 2008 at 12:15 pm
Elisabeth Rédei (11/3/2008)
It should only be your logins unless someone put user objects in there (which is not normally not the case)./Elisabeth
Good deal.
So if I can get the MDF and LDF from last night, could I just put them in the DATA directory after renaming master.mdf and mastlog.ldf and start the service?
That seems crude; I know SQL tends to not like that sort of simple copypasta.
November 3, 2008 at 12:18 pm
... and as soon as things are up and running again; run DBCC CHECKDB on ALL your databases.
Lets' hope not, but if you are unable to restore your backup (in case it is corrupted as well) you rebuild it from your setup CD but of course then you will loose all your logins.
/Elisabeth
elisabeth@sqlserverland.com
MCITP | MCT
http://sqlblog.com/blogs/elisabeth_redei/
http://linkedin.com/in/elisabethredei
November 3, 2008 at 12:19 pm
I read that article...
I have the rebuild master utility up but I've yet hesitated to point it to the mdf, ldf files.
Should I point it to the corrupt master.mdf, mastlog.ldf files, let it rebuild and (hopefully) bring it to enough of an acceptable state to start, then restore master from Enterprise Manager?
Sorry for all the questions, I really appreciate all the quick replies!
November 3, 2008 at 12:25 pm
Elisabeth Rédei (11/3/2008)
... and as soon as things are up and running again; run DBCC CHECKDB on ALL your databases.
Lets' hope not, but if you are unable to restore your backup (in case it is corrupted as well) you rebuild it from your setup CD but of course then you will loose all your logins.
/Elisabeth
This is where my confusion comes into play in terms of what we are considering a backup.
I think their IT people have the mdf/ldf files from last night/this morning, and not a .bak...
If I can't start SQL I can't to a "Restore" anyways though right? Is the recommendation to simply copy those files into the data directory and try to start SQL?
November 3, 2008 at 12:33 pm
You point it to the DATA directory of the Setup CD.
elisabeth@sqlserverland.com
MCITP | MCT
http://sqlblog.com/blogs/elisabeth_redei/
http://linkedin.com/in/elisabethredei
November 3, 2008 at 1:01 pm
OK, after backing up the current system databases, I restored from the CD using that utility and was able to start SQL (THANKS!).
I have 2 databases to attach; I attached the first one successfully.
The second database gives me this frightening-*** error:
Error 3624:
Location: p:\sqltdbms\storeng\drs\include\record.inl:1447
Expression: m_SizeRec > 0 && m_SizeRec <= MAXDATAROW
SPID: 51
Process ID: 4744
:crazy:
Viewing 15 posts - 1 through 15 (of 25 total)
You must be logged in to reply to this topic. Login to reply