SQL*NET problems -- (TNS 510 etc)

We have 2 Vignette servers running on Solaris boxes that access our e-commerce
database hosted on an RS6000 running AIX. We experienced some load balancing
problems on the Solaris boxes, and the Vignette admins claim that it's not entirely hardware related, that there appears to be some interface issues as well. They analyzed the packets bouncing between the app servers and the database and stated that each frame going from the app server to the database
server is repeated 3 times, and after a short wait, 3 more times, followed by
something in SQL called OCANCEL.

Checking the listener.log on the database server revealed no errors, but the sqlnet.log on the app servers had a stream of TNS-12540 and TNS-510's:

12540, 00000, "TNS:internal limit restriction exceeded"
// *Cause: Too many TNS connections open simultaneously.
// *Action: Wait for connections to close and re-try.

00510, 00000, "Internal limit restriction exceeded"
// *Cause: Too many files or sockets open simultaneously (or some other
// resource has been depleted).
// *Action: For further details, trace the operation for protocol
// details.

If this was a resource problem on the database server, wouldn't I see similar
errors in the listener.log on that box?? Could it possibly be a resource issue
on the Solaris boxes?? I'm not entirely sure where to begin looking, I've checked Metalink and all messages referencing similar error messages indicated
that these messages appeared in the listener.log. The database box looks good
actually, plenty of memory and isn't being taxed much.

Thanks,
Jeff T.
thomasje_at_tce.com                                                                      Received on Wed Nov 01 2000 - 11:19:43 CST