My tempdb is eating 1.27 gb of space of my d dive, and now only 10 mb is left in d drive . for this i stoped and started sqlserver but it didnt release much of space.Pls let me know is there any other way so that i can release some space from my tempdb.
Go in to EM and right-click on TempDB. Select properties. Set an upper limit on file growth so that it can't eat more than 80% of your available drive space (taking into consideration other files that may be co-located on the same volume).
It is possible that the above steps will not reset the size of Temp DB, but it should.
That does seem a bit extreme. You should be able to issue a
DBCC ShrinkFile (n, s)
command, where n is the fileid you want to shrink, and s is the target size. Before you do this, check to see if there are any open transactions holding tempdb at this size (issue dbcc opentran in tempdb). If no transactions are holding up space, then DBCC SHRINKFILE should deal with the immediate problem. Beware of tempdb growing again. You should probably set some limits on the filegrowth, then look for the root cause. Especially if this was an "overnight" growth of tempdb, which it sounds like it was.
Yes, perhaps it is a bit extreme; it should be used as a last resort.
However, TempDB is recreated every time you start SQL server. When you issue DBCC SHRINKFILE, I believe that much of the activity takes place on TempDB. I've had varying results with SHRINKFILE (I don't much use it anymore for my large databases).
Given the very small amount of space available to him on his TempDB drive, I thought this might bring about a faster resolution.
Shrinkfile is an incredible pig if you need to shrink user databases. It is much better at just freeing up empty space. If you need to shrink a user database .MDF file, and you have data at the end of the file, SQL Server has to relocate that data somewhere. This causes indexes to be modified, transaction logs to grow, and probably puts a few locks on the tables being moved, just for good measure. Usually, I find it easier to move a few objects to a new filegroup, then do dbcc shrinkfile (n, truncateonly). Then bring back the objects from the new filegroup. Painful, but quicker, I think.