Oracle 8.1.5 & HPUX 11.11. External Proc losing RPC connection

[Phil Daws_2]

Hi:

We have a urgent problem where Oracle is losing a RPC connection under UX 11.11. The message we are getting is "ORA-28576: lost RPC connection to external procedure agent". If we migrate the instance to a UX 11.00 server all works fine. I have checked MetaLink and the UX knowledge base but cannot find anything. Any ideas? or we shall have to downgrade to UX 11.00 :(

[Steven Gillard_2]

Here's the Oracle documentation on the error:

$ oerr ora 28576
28576, 00000, "lost RPC connection to external procedure agent"
// *Cause: A fatal error occurred in either an RPC network connection,
// the extproc agent, or the invoked 3GL after communication had
// been established successfully.
// *Action: First check the 3GL code you are invoking; the most likely
// cause of this error is abnormal termination of the
// invoked "C" routine. If this is not the case, check for
// network problems. Correct the problem if you find it. If all
// components appear to be normal but the problem persists, the
// problem could be an internal logic error in the RPC transfer
// code. Contact your customer support representative.
//

Check the following:

- That the listener is running and is configured to accept calls to external procedures.

- Look in the listeners log file for any more specific errors there.

- Have you configured any of your own or other third party libraries that are called as external procedures? The DBA_LIBRARIES view may help you identify these. Is it one of yours that is failing?

Failing that, follow the oerr instructions and report the problem to Oracle!

Regards,
Steve

[Phil Daws_2]

Nothing out of the ordinary in the listener logfile. It is was of our own libraries. If we run the external proc via the command line using a test harness it works fine. It sometimes works okay via Oracle but 1 in 3 fail. The same tests on UX 11.00 work fine everytime.

[Steven Gillard_2]

OK, I'm no expert on external procedures in Oracle but my basic understanding of how they work is:

- oracle server makes connection to the listener via IPC
- listener spawns extproc process
- extproc process loads the library and makes the request.

Do you know if your procedure is being invoked when it fails or is it failing before that?

My next step would be to run a tusc -f on the listener and then get it to fail. Then look for any system call failures that may cause the problem. Especially the fork / exec calls for extproc and mmap calls to load your library. This may highlight any system problems that cause the error.

I would also recommend you open an Oracle call on this, as you seem to have eliminated your code as the cause of the problem.

Regards,
Steve