|
HI All, I have a database on the mdf file is not increasing but the ldf file increasing rapidly. I have truncated the log file and make it to 1 MB in size but in just one day it baecame 50 GB in size, i take T log twice a day from the maintanece plan. I think from maintanence plan the log file do not truncate while taking backup. Do i need to take more frequently backup.. Thanks Basit Khan
(comments are locked)
|
|
You must have a great deal of transactions updating and or changing data consistently. You should do much more frequent backups. I would start with either 15 or 30 minutes. If you are already doing twice a day but it is growing to 50 GB in size, you need much more frequent backups. You should also read Kimberly's blog on virtual log files. You are probably having problems with tons of virtual log files and aren't even aware of it. What about an uncommitted transaction?
Aug 12 '12 at 09:43 AM
ThomasRushton ♦
Basit, did this response help? Have you made any changes?
Aug 13 '12 at 08:58 PM
Tim
You should also consider if you are doing nightly maintenance that could be growing the log as well.
Aug 15 '12 at 02:12 PM
Tim
Especially with a "Drag&Drop" Maintenance Plan. Starting with a checkdb, running backups, shrinking files and then running index maintenance.... But I'd still also investigate what is causing the log to grow. If by the end of the day SQL Server grew the file to 50GB again, then the server potentially needed the space. Alltogether an overall look can help to understand the cogs involved
Aug 15 '12 at 02:23 PM
DirkHondong
(comments are locked)
|
|
Hi Basit, I think you should start some investigation of what is happening. Gail Shaw wrote a great article on ssc. Take a look: http://www.sqlservercentral.com/articles/Transaction+Logs/72488/ You should keep in mind that shrinking the log isn't best practice at all and should only be the last solution if there's no other way to solve a space problem. Regards Dirk
(comments are locked)
|
|
It's hard to tell if 50GB transaction log is "large" or "small". It depends on your size of disks, your speed of disks, size of the data file(s) etc. Use DBCC SQLPERF(logspace) to find out how much of the transaction log is actually used. If the transaction log is nearly full, you won't be able to shrink it either. (But you probably don't want to do that anyway) If what's happening is that you have a large transaction log, back it up, shrink it and then it grows back to 50GB again in one day, I'd say you should either back it up more frequently or simply let it be 50GB. Otherwise your queries will perform poorly because the transaction log needs to auto grow all the time. (Plus you'll get a lot of virtual files, like Tim said).
(comments are locked)
|
|
check the recovery model of your database by using this command in ssms select recovery_model_desc,name from sys.databases for best practive to shrink a database, please cehck my blog http://sqlservr.blog.com/2012/06/26/best-practice-to-shrink-in-logldf-file-in-sql-server/ if your databse is in simple recovery mode then use USE [dbname] GO DBCC SHRINKFILE (N'dbname_log' , 0) to check for open transaction use dbcc opentran
(comments are locked)
|


Do you have any index maintenance jobs configured against the database? This can cause a great deal of log activity.
Did you get this fixed?