November 22, 2004 at 8:05 am
Good Morning All! Over this past weekend I had an issue I had not come across before and was looking to see anyone else had and if there is anything that can be done to prevent it.
Here is some background. We’re on SQL 2000 sp3. The database sits on a three node cluster, two active one passive. It is a highly transactional database with several hundred users, but lucky for us this happened on a Saturday morning where the user load is about 25%. It was about 63 GB before the trouble. Auto Grow is on.
Saturday Morning it wanted to grow. The disk space was taken from the operating system and the data file grew. However, the database did not recognize the new space. Select transactions went through but any inserts just hung until they timed out. What makes it more fun is the lack of evidence that anything was ever wrong. There were no entries in the SQL server log, and none in the event viewer. After the physical growth of the data file ended but the trouble persisted I ran a DBCC checkdb. The database also grew again. At some point after both the DBCC and the second growth the database went back to normal operation. My gut tells me that the newalloc section of the checkdb was the trigger for the return to normalcy.
So does anyone have any idea why the database would not recognize the new space?
November 22, 2004 at 9:18 am
I assume the space was already allocated on your SAN drive and SQL just picked it up? How much space was added v how much was there? Are you sure the space was completely zeroed out when you were checking?
November 22, 2004 at 9:23 am
Steve,
Yep, the SAN had plenty of space. The Drive the data file is located on is 199 GB. And still, after the second growth, has 59 GB left. It added about 6 GB the first time and about 7 the second (10%) each growth.
Chuck
Viewing 3 posts - 1 through 2 (of 2 total)
You must be logged in to reply to this topic. Login to reply