How to get a full trace of a prog's execution?






I'm trying to exterminate a bug that shows up very sporadically
(but with serious results). The program in question is a CGI script
and I'm having a very hard time rooting out the bug.

It would be extremely useful to be able to toggle on something akin
to "Trace Mode" in the Perl debugger, sending the trace to some
user-specified file. (Since the bug shows up sporadically and
unpredictable, it is not practical to try to find it by using the
Perl debugger.) Is there some other way to get such Trace output?

TIA!

kj
--
NOTE: In my address everything before the first period is backwards;
and the last period, and everything after it, should be discarded.
.



Relevant Pages

  • Re: 2.6.19-rc2-mm1
    ... testing NMI watchdog ... ... The x86_64 nmi watchdog handling looks rather complex. ... I don't think trace all cpus ... BUG: ...
    (Linux-Kernel)
  • [BUG] 2.6.27-rc1 in ext3_find_entry
    ... in the middle of running git; I forget exactly what I was doing. ... BUG: unable to handle kernel paging request at ffff88004fc00006 ... Call Trace: ...
    (Linux-Kernel)
  • Re: cpu load balancing problem on smp
    ... this is not an irq distribution that irqbalance should give ... since this only happens when he adds more iptables rules, ... BUG: ... Call Trace: ...
    (Linux-Kernel)
  • Re: [PATCH] fix cpufreq_stats attrs removal
    ... rmmod cpufreq_stats: Ooops! ... Call Trace: ... The attached patch seems to solve both symptoms and yes... ... BUG: unable to handle kernel paging request at virtual address 6b6b6b6f ...
    (Linux-Kernel)
  • Re: NFS oops in 2.6.26rc4
    ... have a look at the actual filehandle the server is producing. ... BUG: unable to handle kernel paging request at 5a5a5a5a ... Call Trace: ...
    (Linux-Kernel)