Re: Have you ever had an ORA-600 [105] Error when tracing

Michael,

Sounds an interesting problem. Nothing worse than a problem that is not predictable.

I have some spare time tomorrow, so I might to reproduce this on 8.1.6/8.1.7 test instances I have. Do you by any chance have any of the ora-600 trace files that show the full stack trace, to ensure, if I can reproduce it, that it is being caused by the same thing?

Thanks,
Samantha

Michael Brown wrote:

> This particular error has had a bug logged on it for HP-UX under 7.3
> and NT under 8.0. In both cases the bug was closed without
> resolution. I got the error when starting a trace on a user process
> and on the background processes. On more than one occasion/more than
> one instance I have performed traces in this manner (starting and
> stoppping the trace successfully). On one occasion, I got the ORA-600
> [105] when starting the trace and yes down went my database.
>
> I am currently trying to do some research on this issue to see how
> common it is across several variables (in general, to OS, to Oracle
> version, to OS/Oracle version pair, etc.).
>
> Both of the bugs had to do with stopping a trace as opposed to
> starting the trace. That is why I asked what was happening at the
> time of the error.
>
> Based upon what I found on Metalink, I think that most of these errors
> are dropped after the user finds that it was caused by an
> uninitialized memory heap when changing sql trace levels and that the
> error will not cause data corruption. They learn that it is basically
> an infrequent semi-random event tied to doing an sql trace and drop
> it. This leaves Oracle under no pressure to fix the problem.
>
> I am especially interested in anyone who has received this error under
> 8.1.6 or 8.1.7 since that would indicate the issue is still
> unresolved.
>
> Mike
>
> On Mon, 11 Jun 2001 13:27:54 -0600, Samantha Hall
> <samanthahall_at_qwest.net> wrote:
>
> >Michael,
> >
> >Would be a good idea if you can give more information on this. Are you
> >looking for info because you have seen this error and need to diagnose
> >it ?
> >
> >Internal ORA-600's can be very specific to database version and code
> >that is being executed at the time to cause the error.
> >
> >For better help if you have seen this error, Oracle Support/Metalink is
> >a good start, otherwise you need to supply the same information you are
> >requesting- database version, platform/OS, what you were doing at the
> >time of the error.
> >
> >BTW, I think is error can be reported when you have set a system tracing
> >event for a background process and then try ro unset it. Not a good
> >idea! Easy to test and bring the instance down :-)
> >
> >Regards,
> >Samantha
> >
> >
> >Michael Brown wrote:
> >
> >> I am trying to collect some data about the frequency of this error
> >> occuring. If you have ever had a database go down with an ORA-600
> >> [105], please let me know as much as you can of the circumstances
> >> under which the error occured including how you activated the trace,
> >> what you were tracing (user or specific background process), what
> >> hardware/OS platform and what database version.
> >>
> >> Thanks for your help.
> >>
> >> Michael Brown
> >> Sr. DBA
> >> Glen Raven, Inc.
> >> (336)586-1146
> >> mbrown_at_glenraven.com
> >
Received on Mon Jun 11 2001 - 22:59:23 CDT