Oracle FAQ | Your Portal to the Oracle Knowledge Grid |
![]() |
![]() |
Home -> Community -> Usenet -> c.d.o.server -> Oracle80 Process Memory Leak
Need help with the Oracle80 Process Memory Leak.
Here is my situation:
System: HP Netserver with NT4.0 Sp5, 2GB Memory and 4 Processor. Oracle Database: SGA
db_block_size = 4096 db_block_buffers = 51200 (200MB) share_pool_size = 314572800 (300MB) sort_area_size = 262144 log_checkpoint_interval = 8000 processes = 500 parallel_max_server = 16 log_buffer = 61440 sequence_cache_ebtries = 100 sequence_cache_has_buckets = 100
I am running an a third party application called Picis (healthcare monitoring aplication) using a third party ODBC driver call INTERSOLVE 3.11 32-Bit Oracle8. I'm running Oracle 8.05 with the latest Oracle patches on it.
My problem is that the Oracle80 process virtual bytes keeps increasing
and when I get to about 1.7 to 1.8 GB of virtual bytes, I get a TNS
Listener error ORA-12500 and no more connection can be made to the
database. The only way to resolve this is to reboot the database
server
and it will last me about 3 days of heavy usage before I have to reboot
it again.
I've been monitoring the Oracle80 process virtual bytes and the situation is this, every times I made connection to the database, it uses certain of the virtual bytes memory and when I close the application, the connection drop, but the virtual bytes memory does not
drop back to where I first started the application as it suppose to.
It
actually keeps between 40 to 50 % of the memory it uses and that's why
the virtual bytes keeps on increasing. The problem is not just
localize
to the third party application (Picis) that I'm using, the problem
occur
also when I'm using Oracle tools such Schema Manager, Instance Manager,
etc... It uses certain amount of virtual bytes on the Oracle80 process
when it connected to the database, but it does not give it all back
when
I closes them.
I've been working with Oracle and the third party vendor and so far
I've
not been able to reslove anything. Oracle does not have an answer for
and they ask us to apply the latest patch and that did not solve the
problem. The vendor think that our database server is over size with
2GB of Memory on NT server 4.0. We do not run NT Enmterprise server at
our location. I'm stuck and unable to resolve this problem at the moment. Does any body ever ran into anything similar and able to reolsve it or might have a solution for me.
Thanks,
Boran Received on Mon Jan 17 2000 - 16:36:57 CST
![]() |
![]() |