[Orca-dev] [PATCH] Updates to include additional Solaris performancemetrics for graphing

[Blair Zajac]

Sean O'Neill wrote:
> 
> At 07:58 PM 12/18/2002 -0800, Blair Zajac wrote:
> 
> >Hi Sean,
> >
> >Could you do me a favor.  The patch you sent got munged somewhere along
> >the way (line wrapping) and I've also applied Liston's patch to orcallator.se
> >and orcallator.cfg.in.
> >
> >Could you update your Orca working copy (with your modifications in it)
> >to HEAD and the latest 'svn diff' output as an attachment.  That way it
> >should be better protected.
> 
> K svn repos updated to HEAD (rev172 at the time) with my updates merged
> in.  HEAD-svn.diff is file attachment with svn diff output.

Hi Sean,

Thanks.  This one applied cleanly.

Do you have sample plots for these?  I'd like to add them to the

    http://www.orcaware.com/orca/docs/orcallator.html

page.  This will then let us use href elements for the new plots so
people can get more information on it.

Rather lengthy review below.

Also, please include the log message with any diffs.  It makes it
easier to have the entire patch and log message in one email,
rather than finding the old email with the log message.

Regarding the new measurements, let's have all measurements that
are something per second have a "/s" in the name.  I haven't looked
closely at which ones are, but I may note some below.

I definitely want to get these new column names right so we don't
have to change them later and mess up plots.

In orcallator.cfg.in:

 plot {
+title                  %g Interrupts, System Calls, Traps
+source                 orcallator
+data                   interrupts
+data                   intrthread
+data                   system_calls
+data                   trap

How about we rename 'trap' to 'traps'.

Are these per second?

+legend                 Interrupts/s
+legend                 ThreadInterrupts/s
+legend                 ThreadInterrupts/s

Missing spaces here in these legends.

+title                  %g Forks, Vforks, & Execs

There's an extra , before the &.

+data                   forks
+data                   vforks
+data                   execs
...
+legend                 forks/s
+legend                 vforks/s
+legend                 execs/s

Is this data is already per second, or the total number in a measured
interval?

+legend                 Voluntary Switches/s
+legend                 Involuntary Switches/s
+legend                 Total

Same here, /s or not.

+title                  %g namei/s, iget/s, dirblk/s

Can we make this title more useful?  I don't know what these mean
without looking them up.

+title                  %g Pageins/s, Pageouts/s, PagesFreed/s

How about "Pageins/s, Pageouts/s & PagesFreed/s"

+data                   pgin
+data                   pgout
+data                   dfree

/s or not???

+title                  %g Pages PagedIn/s & Pages PagedOut/s

How does this data set differ from the Pageins/s, Pageouts/s,
PagesFreed/s in the previous plot?

+title                  %g Major & Minor & Protection Page Faults

Change to "%g Major, Minor & Protection Page Faults"

These also look like per second.

+title                  %g COW & ZFOD Page Faults

Can we make this longer "Copy-on-write & Zero-fill-on-demand page
Faults"?

I noticed in orcallator.se that ufsinopage is measured, but there's
no plot for it.


Now in orcallator.se:

+// Version 1.37:   Dec 20, 2002 Additions by Sean O'Neill - Happy Holidays !!
+//                              Added several new variables for Orca to collect
+//                              and graph including:
+// pgrec - page reclaims (include pageout)
+// pgfrec - page reclaims from free list

Anal, but can the list of names be justified as all the other comments are?

+// ufsinopage - inodes taked with no attached pages

Typo on taked?


In lib/Makefile.in and ./orcallator/start_orcallator.sh.in, how
about we put the 3.2.1 and 3.3 directories in a SE directory,
so it looks like

lib/SE/3.2.1
lib/SE/3.3

and that it installs this way also.

Regarding ./lib/Orca/Constants.pm, I'd like to keep this change out of
this patch.  How do the output PNG files look with this change?

One of the things that these numbers are designed for is that the number
of data points in a plot is less than the number of pixels in a plot,
otherwise RRD needs to resample the data, and this change from 1.5
days to 3 days bumps the number of data points from 432 to 864, which
is greater than the default width of 500.

Best,
Blair

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