Shrinking the log file does not reduce size

Question

I have a database which has a 350 MB data file (.mdf) and a 4.9 GB log file (.ldf). The recovery model is set to FULL.

When I try to shrink the log file, it's not shrinking.

I know shrinking a database is not good and it should not be done. But still i am trying to do it for shrinking the log file.

When I ran

DBCC SQLPerf(logspace) 

I found that the log size is 4932 MB and Log space used is 98.76%!

Then I tried this command

USE <databasename>;
DBCC loginfo;

Now almost all VLFs are "status 2" which means all are in use.

I tried to take a log backup and then shrink the log file. Shrinking didn't reduce the size.

I changed the recovery model to SIMPLE and tried shrinking again, but this also didn't help.

I checked for open transactions

DBCC opentran (database);

and found that no transaction is open now.

What is stopping me from shrinking the log file? How can I solve this?

Answer 1

Here is the answer to my own question.

Run the below query to get information about the log file's reuse wait:

SELECT log_reuse_wait_desc
FROM sys.databases
WHERE name = 'DBName'

I got the following output:

log_reuse_wait_desc
-------------------
REPLICATION 

There were some replication-related objects remaining in the database, even after removing the replication.

To remove the replication from the database, sp_removedbreplication can be used. But it didn't work for us as replication was not active at the time and actually replication had been removed long before.

The solution was to import the database contents to another database using the import option of SQL Server.

Answer 2

Read How to Shrink SQL Server log for an explanation how the circular nature of the log may prevent shrink after truncation. Is possible that you log's last LSN point into a VLF that is at the tail of the LDF. Counter intuitively you must advance the log, by generating log writes, to allow it to shrink.

Answer 3

Steps for shrinking the log are going to be:

Backup transaction log through either SSMS or T-SQL and then perform a shrink.

commands for SSMS are under the tasks if you right click the database name.

BACKUP LOG <Databasename> TO DISK = N'<path\database_log.ldf';
GO

DBCC SHRINKFILE (<FileName>, <TargetSize>) WITH NO_INFOMSGS

You will probably have to do this multiple times.

If there is a transaction or job blocking the action, use Activity monitor to identify the process and kill it, or use the SQL Agent job activity monitor to end the job.

source: http://support.microsoft.com/kb/907511

Answer 4

I have found that I have to perform 2 or 3 backups of both the database and the transaction log to get the transaction log to actually reduce in size. I have a database that was created with Full recovery model. Every night it performs backups of the database and the transaction log but inevitably the transaction log seems to continually grow over 2-3 weeks. When the remaining disk space gets to 1GB I will see that the transaction log is about 30GB. I followed the steps recommended by Microsoft and after the 4th or 5th iteration of backing up both the database and the transaction log the transaction log will finally release its extra space and shrink. Then I go back and delete the multiple backups I have created.

Answer 5

You need to create a backup first, dependent on the backup model that is set up for the database before you can shrink the database.

You can try running this:

USE <databasename>
GO

BACKUP DATABASE <databasename> TO DISK '<absolute path goes here>\<databasename>.bak';
GO

Or you can do that from SSMS and use the graphical tools available (see here for details: http://msdn.microsoft.com/en-us/library/ms187510.aspx)

Once you've backed up your database you can compress it. However, shrinking the database is not a good idea since heavy index fragmentation will occour and searching for data will become slow.

Hope this helps.

Answer 6

My work around for the Replication that is blocking shrinking log file is:

1. Set DB Recovery Model to Simple
2. Take DB offline
3. Create backup of log file (just in case)
4. Delete log file
5. Bring DB online

In my case it worked. After bringing DB online log was created automatically and it's size was 512kb instead of 70GB. But this is only a workaround. The root problem is not resolved. In my case we are using replication.