| why db need kill session instantly [message #500551] |
Tue, 22 March 2011 03:52  |
runinwind
Messages: 7
Registered: February 2005
|
Junior Member |
|
|
Hello all, i have a 10G2 DB server running on RedHat Enterprisse. When the session count over 500, it will run very slow, mainly library latch waiting occuring, and the listener will stop accidentally. Restart the listener and same things happen.
If i kill some session using:
select 'kill -9 ' || p.spid from v$session s,v$process p where s.paddr=p.addr and s.program = 'AppServer.exe' and last_call_et > 30*60 and status = 'INACTIVE'
This will typically kill 200 sessions, the server will receover and runs well.
The attached pic shows that kill session at 3:10, before and after the server performance different very much.
I just wonder what's the wrong with this server? When it slow, i tried flush shared_pool but can't resolve.
-
Attachment: db_0322.jpg
(Size: 100.08KB, Downloaded 1678 times)
|
|
|
|
|
|
|
|
] 
Blog
Search
Help
Members
Register
Login
Home
