SQL Server Log Errors
Q. What do these error msgs mean, why are they showing up and how do I fix them?
Errors:
1. Autogrow of file 'fl1stDwProd2_Log' in database 'fl1stDwProd2' took 105468 milliseconds. Consider using ALTER DATABASE to set a smaller FILEGROWTH for this file.
2. WARNING: EC 6ba3b548, 0 waited 300 sec. on latch 42c862c8, class MISC. Not a BUF latch.
System setup:
Db space allocation: 11799MB
Log space allocation: 49497MB
Db Size: 12,081,344K
Lob Size: 50,684,544
Both Auto Grow by 10% Unrestricted, Recovery Model Full, Auto Shrink.
Full backup nightly
Software/Hardware used:
Software/Hardware used:
ASKED:
January 10, 2006 10:11 AM
UPDATED:
January 11, 2006 10:32 AM
Answer Wiki:
I do believe this is telling you that when the files get full, it is taking a long time to execute the AUTOEXTEND feature. We try to set the AUTOEXTEND feature to a set amount i.e we will set it 512 MB instead of setting it to a percentage. When set at a percentage of the file size, you have no control of how big the newly extended amout is or how long it will take (due to the fact that the percentage gets LARGER with each extent.) Remember, when this happens, the database will slow down to a crawl, so it is important to know how your system will react to such activities.
To see all answers submitted to the Answer Wiki: View Answer History.
I’ve backed up the tran log and resized.
The size is now smaller than the db and I restriced the growth.
Thanks so much!
Be careful when you restrict the growth. If someone performs a large transaction (for example delete a million rows from a table), the log file could fill up and can’t extend. As a result, SQL Server will halt. The important thing is to perform backups on the transaction log so the space can be reused. If you leave auto extend on and you backup the log, the log file will only be as big as needed to perform the largest transaction ever done. If someone performs an unusually large transaction, you could shrink the file again.