[Orca-users] Re: Segmentation Fault with 1.35

[Blair Zajac]

Alan,

Thanks.  This will be in 1.36.

On another note, I want to add a new measurement suggested by
John Langner on measuring the average disk service time.  For
the non-RAWDISK code, this will add another measurement

    put_output(sprintf("disk_svct_%s", GLOBAL_disk[i].info.long_name),
    sprintf("%16.5f", GLOBAL_disk[i].service));

What would we use for the RAWDISK code?

    put_output(sprintf("disk_runp_%s", RAW_disk[i].long_name),
               sprintf("%16.5f", RAW_disk[i].servce));

Here's John's original message:

> I find the svc_t column from "iostat -x" more interesting than
> the other numbers.  It seems that this would be the most revealing
> number about how much applications are slowed down waiting
> for file access.
>
> One of the Sun World disk performance columns pointed out the
> non-obvious fact that a mildly busy disk can actually produce shorter
> average service times than one very lightly loaded.
>
> Adding a disk average service time plot turned out to be very simple.
>
> ..../orca/lib/orcallator.se  --  Copy the two lines containing "disk_runp"
>     and make a slight change:
>
>     put_output(sprintf("disk_svct_%s", GLOBAL_disk[i].info.long_name),
>     sprintf("%16.5f", GLOBAL_disk[i].service));
>
> ..../orca/lib/orcallator.cfg - Copy the disk run percent section and make
>     a few changes:

Best,
Blair

-- 
Blair Zajac <blair at orcaware.com>
Web and OS performance plots - http://www.orcaware.com/orca/