Urgent: Latch free increasing on latch#94

Hi all,

I am having issue right now on Production database. There are total latch free event show 16 latch free events and its increasing.

It is having 94 in P2 column of v$session_wait, which is 'temp lob duration state obj allocation' event in v$latch.

There is no note on metalink on this latch name.

Also, there is no row in v$latchholder.

What should I do? Its very urgent.

Thanks in adv for any help.

Update: Now even ORACLE.exe (type Background) is also waiting on this latch and I got foll. msg in PMON trace.

*** 2003-07-17 09:18:09.421
*** SESSION ID:(1.1) 2003-07-17 09:18:09.390
PMON unable to acquire latch 13cd3b8 temp lob duration state obj allocation
*** 2003-07-17 09:18:09.421
*** 2003-07-17 09:28:10.765
PMON unable to acquire latch 13cd3b8 temp lob duration state obj allocation
*** 2003-07-17 09:28:10.765
*** 2003-07-17 09:38:12.109
PMON unable to acquire latch 13cd3b8 temp lob duration state obj allocation
*** 2003-07-17 09:38:12.109
*** 2003-07-17 09:48:13.453
PMON unable to acquire latch 13cd3b8 temp lob duration state obj allocation
*** 2003-07-17 09:48:13.453
*** 2003-07-17 09:58:14.765
PMON unable to acquire latch 13cd3b8 temp lob duration state obj allocation
*** 2003-07-17 09:58:14.765
*** 2003-07-17 10:08:17.109
PMON unable to acquire latch 13cd3b8 temp lob duration state obj allocation
*** 2003-07-17 10:08:17.109

Sam,

Please check out alert log file for any other error messages associated with the core dump file.

I have noticed on one of the instances on my production, the growth of the tablespace xyz_lob was so enormerous over an year period, but seem to have few megs of data. So the query issued against this fragmented LOB tablespace started to hold latches. The solution was to export- drop and recreate the tablespace - import. This fixed the problem.

Check whether you have any such issues...

Sam