[sql] The transaction log for the database is full

I have a long running process that holds open a transaction for the full duration.

I have no control over the way this is executed.

Because a transaction is held open for the full duration, when the transaction log fills, SQL Server cannot increase the size of the log file.

So the process fails with the error "The transaction log for database 'xxx' is full".

I have attempted to prevent this by increasing the size of the transaction log file in the database properties, but I get the same error.

Not sure what I should try next. The process runs for several hours so it's not easy to play trial and error.

Any ideas?

If anyone is interested, the process is an organisation import in Microsoft Dynamics CRM 4.0.

There is plenty of disk space, we have the log in simple logging mode and have backed up the log prior to kicking off the process.

-=-=-=-=- UPDATE -=-=-=-=-

Thanks all for the comments so far. The following is what led me to believe that the log would not grow due to the open transaction:

I am getting the following error...

Import Organization (Name=xxx, Id=560d04e7-98ed-e211-9759-0050569d6d39) failed with Exception:
System.Data.SqlClient.SqlException: The transaction log for database 'xxx' is full. To find out why space in the log cannot be reused, see the log_reuse_wait_desc column in sys.databases

So following that advice I went to "log_reuse_wait_desc column in sys.databases" and it held the value "ACTIVE_TRANSACTION".

According to Microsoft: http://msdn.microsoft.com/en-us/library/ms345414(v=sql.105).aspx

That means the following:

A transaction is active (all recovery models). • A long-running transaction might exist at the start of the log backup. In this case, freeing the space might require another log backup. For more information, see "Long-Running Active Transactions," later in this topic.

• A transaction is deferred (SQL Server 2005 Enterprise Edition and later versions only). A deferred transaction is effectively an active transaction whose rollback is blocked because of some unavailable resource. For information about the causes of deferred transactions and how to move them out of the deferred state, see Deferred Transactions.

Have I misunderstood something?

-=-=-=- UPDATE 2 -=-=-=-

Just kicked off the process with initial log file size set to 30GB. This will take a couple of hours to complete.

-=-=-=- Final UPDATE -=-=-=-

The issue was actually caused by the log file consuming all available disk space. In the last attempt I freed up 120GB and it still used all of it and ultimately failed.

I didn't realise this was happening previously because when the process was running overnight, it was rolling back on failure. This time I was able to check the log file size before the rollback.

Thanks all for your input.

I met the error: "The transaction log for database '...' is full due to 'ACTIVE_TRANSACTION' while deleting old rows from tables of my database for freeing disk space. I realized that this error would occur if the number of rows to be deleted was bigger than 1000000 in my case. So instead of using 1 DELETE statement, i divided the delete task by using DELETE TOP (1000000).... statement.

For example:

instead of using this statement:

DELETE FROM Vt30 WHERE Rt < DATEADD(YEAR, -1, GETDATE())

using following statement repeatedly:

DELETE TOP(1000000) FROM Vt30 WHERE Rt < DATEADD(YEAR, -1, GETDATE())

Try this:

USE YourDB;  
GO  
-- Truncate the log by changing the database recovery model to SIMPLE.  
ALTER DATABASE YourDB
SET RECOVERY SIMPLE;  
GO  
-- Shrink the truncated log file to 50 MB.  
DBCC SHRINKFILE (YourDB_log, 50);  
GO  
-- Reset the database recovery model.  
ALTER DATABASE YourDB
SET RECOVERY FULL;  
GO 

I hope it helps.

To fix this problem, change Recovery Model to Simple then Shrink Files Log

1. Database Properties > Options > Recovery Model > Simple

2. Database Tasks > Shrink > Files > Log

Done.

Then check your db log file size at Database Properties > Files > Database Files > Path

To check full sql server log: open Log File Viewer at SSMS > Database > Management > SQL Server Logs > Current

My problem solved with multiple execute of limited deletes like

Before

DELETE FROM TableName WHERE Condition

After

DELETE TOP(1000) FROM TableName WHERECondition

This is an old school approach, but if you're performing an iterative update or insert operation in SQL, something that runs for a long time, it's a good idea to periodically (programmatically) call "checkpoint". Calling "checkpoint" causes SQL to write to disk all of those memory-only changes (dirty pages, they're called) and items stored in the transaction log. This has the effect of cleaning out your transaction log periodically, thus preventing problems like the one described.

I had this error once and it ended up being the server's hard drive that run out of disk space.

If your database recovery model is full and you didn't have a log backup maintenance plan, you will get this error because the transaction log becomes full due to LOG_BACKUP.

This will prevent any action on this database (e.g. shrink), and the SQL Server Database Engine will raise a 9002 error.

To overcome this behavior I advise you to check this The transaction log for database ‘SharePoint_Config’ is full due to LOG_BACKUP that shows detailed steps to solve the issue.

The answer to the question is not deleting the rows from a table but it is the the tempDB space that is being taken up due to an active transaction. this happens mostly when there is a merge (upsert) is being run where we try to insert update and delete the transactions. The only option is is to make sure the DB is set to simple recovery model and also increase the file to the maximum space (Add an other file group). Although this has its own advantages and disadvantages these are the only options.

The other option that you have is to split the merge(upsert) into two operations. one that does the insert and the other that does the update and delete.

Do you have Enable Autogrowth and Unrestricted File Growth both enabled for the log file? You can edit these via SSMS in "Database Properties > Files"

Try this:

If possible restart the services MSSQLSERVER and SQLSERVERAGENT.

The following will truncate the log.

USE [yourdbname] 
GO

-- TRUNCATE TRANSACTION LOG --
DBCC SHRINKFILE(yourdbname_log, 1)
BACKUP LOG yourdbname WITH TRUNCATE_ONLY
DBCC SHRINKFILE(yourdbname_log, 1)
GO

-- CHECK DATABASE HEALTH --
ALTER FUNCTION [dbo].[checker]() RETURNS int AS BEGIN  RETURN 0 END
GO

Here's my hero code. I've faced this problem. And use this code to fix this.

 USE master;

    SELECT 
        name, log_reuse_wait, log_reuse_wait_desc, is_cdc_enabled 
    FROM 
        sys.databases 
    WHERE 
        name = 'XX_System';

    SELECT DATABASEPROPERTYEX('XX_System', 'IsPublished');


    USE XX_System;
    EXEC sp_repldone null, null, 0,0,1;
    EXEC sp_removedbreplication XX_System;


    DBCC OPENTRAN;
    DBCC SQLPERF(LOGSPACE);
    EXEC sp_replcounters;



    DBCC SQLPERF(LOGSPACE);