G'Day Rick,

On Mon, May 05, 2008 at 11:15:01AM -0500, rickey c weisner wrote:
> Greetings,
> I used the SE toolkit for years. I was particularly fond of
> nx.se. It was not unusual for me to have to add support
> for new NIcs. 
> 
> So I have written my own tool using C and the kstat api.
> It gives you the same output format as nx.se.

Don't take this the wrong way - porting nx.se as a standalone tool is
a good idea... :)

Tim and I have been maintaining nicstat for a number of years which
serves a similar purpose (I created the Perl version of nicstat for the
exact same reason - it was easier to toe around than the SE Toolkit).
nicstat runs on both Solaris and Linux:

http://blogs.sun.com/timc/entry/nicstat_the_solaris_network_monitoring

# nicstat 5
    Time      Int   rKB/s   wKB/s   rPk/s   wPk/s    rAvs    wAvs %Util    Sat
17:05:17      lo0    0.00    0.00    0.00    0.00    0.00    0.00  0.00   0.00
17:05:17  e1000g0    0.61    4.07    4.95    6.63   126.2   628.0  0.04   0.00
17:05:17  e1000g1   225.7   176.2   905.0   922.5   255.4   195.6  0.33   0.00
[...]

The idea was to provide the most concise and generic view of these metrics.
nx.se's output may be ideal as a verbose option that may contain OS specific
metrics.

I've also asked that nicstat be included in Solaris - either as an option
to netstat, dladm or standalone.  Including either nicstat or an nx.se
port would be extreamly valuable.

I think it makes sense for us to discuss this here on perf-discuss, and
see if we can agree what to do.  This will including deciding if it should
be a netstat option, dladm, or stand alone.

In the years that have passed, I think the default output of nicstat can be
improved to have these fields:

    Time      Int   rKB/s   wKB/s   rPk/s   wPk/s    %Util    Sat   Err

ie, drop the average size fields and add an error summary.  I'd like see
utilization, saturation and errors provided as standard metrics for all
resources.  Break downs are necessary too, to understand what they mean.

Brendan

-- 
Brendan
[CA, USA]
_______________________________________________
perf-discuss mailing list
perf-discuss@opensolaris.org

Reply via email to