Re: Informix crash
- From: jprenaut@xxxxxxxxx
- Date: Mon, 30 Jul 2007 07:17:14 -0700
On Jul 29, 6:07 am, "Neil Truby" <neil.tr...@xxxxxxxxxxx> wrote:
IDS 10.0FC5 on AIX 5.3
Don't suppose much can be told from this, but has anyone any ideas about
this failure yesterday?:
Well one of the most important pieces of the AF file doesn't appear to
be working, and that's the stack unwind. It was only able to get to
the signal handler and then it stops. So from just the AF file
there's no way to tell what the thread was doing. Oddly this is the
2nd AF on a 10.00.FCx that I've seen recently where the stack didn't
unwind properly. That really handicaps support in terms of trying to
map this up to a known defect. I just did a quick test of 10.00.FC5
on an AIX box and crashed my server and the AF file did generate the
stack unwind ok, so there might be something "odd" about the stack
that caused the stack unwind code to fail and not produce output
beyond the signal handler frame. The onstat -g stk all output in the
AF file looks ok, so it doesn't look like all stack unwinding. So
unfortunately, without the stack unwind and no current or previously
parsed SQL statements in the AF file, there isn't much to tell other
then the previously mentioned -255 SQL error.
Jacques
.
- References:
- Informix crash
- From: Neil Truby
- Informix crash
- Prev by Date: 2008 IIUG Informix Conference - Call For Presentations
- Next by Date: Re: long transaction, onstat -l, flag A----
- Previous by thread: Re: Informix crash
- Next by thread: Re: PID - Post-Interim Drop (was: Informix 10 permissions problem?)
- Index(es):