Re: "Unexpected oracle server termination..."

In article <CG936G.H4w_at_freenet.carleton.ca> ah513_at_Freenet.carleton.ca (Doug Harris) writes:
>
>In a previous article, ekleindi_at_cs.uml.edu (Edgar Kleindist) says:
>
>> We just installed Oracle7 on a Sun workstation (SunOS 4.1.3).
>>
>> One of my users got this error message when running one of her
>> sql scripts.
>>
>> Unexpected oracle server termination by signal 6
>>
>> Any idea what this error message is complaining about?
>
> The server process blew up. There should be all sorts of interesting
>reading in the server trace file (the alert file will contain the
>OS pid of the culprit). We're running 7.0.12 on SunOS 4.1.3 and get
>this sort of thing now and then. Seems that some of the shared pool
>memory managemet is getting corrupted (UGA heap usually). Try running
>the same statement with a dedicated server (less buggy than the multi-
>threaded server) and maybe it will "go away". Hopefully these bugs will
>be worked out by Oracle "real soon now".
>
> Waiting for a "clean 7.0.15"...
>
> - Doug
>--
While it is true that the server died, the reasons can be many. My guess is that the bug has little to do with MTS.

Contact ORACLE support and log a tar if you still have the trace file (in USER_DUMP_DEST or $ORACLE_HOME/rdbms/log).

By the way, signal 6 is probably not what caused the problem. When ORACLE detects a fault such as segmentation violation, it dumps its stack to the trace file. It then sends itself a signal 6 to cause a core dump.

                                    Thanks,
                                    Rick
                                    Rick Wessman
				    Network Products Division
				    Oracle Corporation
                                    rwessman_at_us.oracle.com