Success! I've now learnt, from our network security admin guy, that the
Firewall Manager on the new network is configured by default to timeout
inactive connections after 60 minutes. Even though all the user clients
are inside the firewall, the server is specially ringfenced because the
Have you considered enabling DCD? This causes oracle to send pseudo-keepalive packets
periodically, which should keep your sessions active from the firewall's perspective.
You enable this by adding
sqlnet.expire_time = ??
to sqlnet.ora on the DB server, where ?? is the frequency (in
--- Norris, Gregory T [ITS]
[EMAIL PROTECTED] wrote:
Have you considered enabling DCD?
Greg,
short of migrating to *nix, dead connection detection
does not work on win32. not in 8.1.7.4, not in
9.2.0.4.
possibly in 10g? ;)
Pd
This causes
oracle to send pseudo-keepalive packets
Paul
Sounds suspiciously like an issue we had when se set up our latest system -
something in the firewall or network set-up has a TCP time-out in it. TCP
(or the network or something I'm no network expert) drops the connection
after 60 minutes resulting in client and Oracle server no longer
Hi Paul,
I'd start with reading Anita Bardeen's paper,
available on the NYOUG site:
http://www.nyoug.org/presentations.htm
ORA-3113's, 600's and 7445's Oh My!
Are these connections local, like on the same subnet,
or is there a firewall or router in between that could
be performing network
No, There is no log generated. Below is the only message I am getting.
hp204:NAMES :/home/oracle svrmgrl
Oracle Server Manager Release 3.1.6.0.0 - Production
Copyright (c) 1997, 1999, Oracle Corporation. All Rights Reserved.
ORA-03113: end-of-file on communication channel
SVRMGR
Unfortunately he's rebooted so the box is probably clean.
Did all of your products relink successfully?
Are there any errors in the OS system logs?
Would any of the patches have changed you kernel settings for shared memory
or semaphores?
Can you run the HP equivalent of truss (I believe it's
Any log file output?
Mike
-Original Message-
[mailto:[EMAIL PROTECTED]
Sent: 06 August 2003 16:29
To: Multiple recipients of list ORACLE-L
Applied 250 OS patches on HP-UX 11.11 server and after that getting
ORA-3113 error message every time I do svrmgrl or sqlplus /nolog connect /as
Shmmax is greater than 4G and oracle 8.1.6 32 will do this.
Set shmmax 4G
Check HP docs for recommended setting.
??
-Original Message-
Sent: Wednesday, August 06, 2003 12:10 PM
To: Multiple recipients of list ORACLE-L
left over memory segments/shared memory/semaphores?
ipcs is your
I tried relinking all and rebooting server then relink again but no luck.
-Original Message-
Sent: Wednesday, August 06, 2003 10:10 AM
To: Multiple recipients of list ORACLE-L
left over memory segments/shared memory/semaphores?
ipcs is your friend, oh and its destructive brother ipcrm
..sounds like a IPC problem.
Check ora network and listener config.
relink client
relink utilities
-Original Message-
[mailto:[EMAIL PROTECTED]
Sent: Wednesday, August 06, 2003 11:49 AM
To: Multiple recipients of list ORACLE-L
No, There is no log generated. Below is the only
left over memory segments/shared memory/semaphores?
ipcs is your friend, oh and its destructive brother ipcrm :)
joe
[EMAIL PROTECTED] wrote:
No, There is no log generated. Below is the only message I am getting.
hp204:NAMES :/home/oracle svrmgrl
Oracle Server Manager Release 3.1.6.0.0 -
Did you relink sqlplus?
Did you relink client?
-Original Message-
[mailto:[EMAIL PROTECTED]
Sent: Wednesday, August 06, 2003 11:29 AM
To: Multiple recipients of list ORACLE-L
Applied 250 OS patches on HP-UX 11.11 server and after that getting
ORA-3113 error message every time I do
Hi Stefania,
AFAIK, there are several bugs, all having to do with problems in processing
the cursor queries. Try searching metalink. Possible solutions are:
- add aliases for every table/view in the from-clause
- don't use parallel query for the cursor query
- don't use functions in the group
14 matches
Mail list logo
