[Orca-users] SE toolkit keeps coring

[Blair Zajac]

Anil wrote:
> 
> My SE toolkit keeps coring. Anyone seen this or have
> pointers to a fix? I hope the formatting in this email
> won't screw up the debug info...  TIA.
...

> RAW_disk[75]._wlastupdate<3881396440502497>)
> if (elapsed_etime<3.00936e+11> > <0>)
> hz_etime = elapsed_etime<3.00936e+11> / nanosecond<1e+09>
> big_etime = <1024> * hz_etime<300.936>
> RAW_disk[75].reads = (_reads<15621769> -
> RAW_disk[75]._reads<15621647>) / hz_etime<300.936>
> Bus Error (core dumped)

Hi Anil,

It's not clear what the problem is.  Here's the piece of code
that appears to core dump:

          elapsed_etime = (_wlastupdate - RAW_disk[rdisk]._wlastupdate);
          if (elapsed_etime > 0)  {
            hz_etime = elapsed_etime / nanosecond;
            big_etime = 1024.0 * hz_etime;
          } else {
            elapsed_etime = nanosecond;
            hz_etime = 1.0;
            big_etime = 1024.0;
          }
          RAW_disk[rdisk].reads  =(_reads-RAW_disk[rdisk]._reads)  /hz_etime;
          RAW_disk[rdisk].kreads =(_nread-RAW_disk[rdisk]._nread)  /big_etime;

There's nothing here that looks like it would core dump.  rdisks is
75, hz_etime is 300.936, big_etime is not zero.  So we're not getting
a divide by zero.

Can you run strace on the process and see what you get here?

Best,
Blair

-- 
Blair Zajac <blair at orcaware.com>
Plots of your system's performance - http://www.orcaware.com/orca/