[Orca-users] Snapshot 480

Cockcroft, Adrian acockcroft at ebay.com
Tue Aug 30 10:42:44 PDT 2005


Thanks Dimitry,

I'm getting old and forgetful :-)

I appreciate your work on this, I need to use it then I'll remember!

Cheers Adrian

-----Original Message-----
From:	Dmitry Berezin [mailto:dberezin at surfside.rutgers.edu]
Sent:	Tue 8/30/2005 5:45 AM
To:	Cockcroft, Adrian; 'Blair Zajac'; 'Nguyen, Binh T'
Cc:	orca-users at orcaware.com
Subject:	RE: [Orca-users] Snapshot 480

Adrian,

You have probably forgotten, but I have used your workload class as a base
for the workinfo class that now comes with orca and is integrated into
orcallator.se. Instructions are included in the workinfo_class.se itself and
I have also explained the usage of it several times on the lists here, so it
should be somewhere in the archives. All it takes now to get that per
process (group) data is setting a couple of variables before starting
orcallator.

Actually, here is the latest discussion of this with Justin Buhler -
http://www.orcaware.com/pipermail/orca-users/2005-June/004371.html

  -Dmitry.

> -----Original Message-----
> From: orca-users-bounces+dberezin=acs.rutgers.edu at orcaware.com
> [mailto:orca-users-bounces+dberezin=acs.rutgers.edu at orcaware.com] On
> Behalf Of Cockcroft, Adrian
> Sent: Tuesday, August 30, 2005 1:57 AM
> To: Blair Zajac; Nguyen, Binh T
> Cc: orca-users at orcaware.com
> Subject: RE: [Orca-users] Snapshot 480
> 
> Years ago I contributed the workollator.se script that con be configured
> to collect microstate based data on any pattern matched group of
> processes, and display the data using orca. It was developed while I was
> working with Justin Buhler on the Olympics systems at Salt Lake and was
> updated and used again in Athens. In particular it was used to monitor
> several Oracle instances and a bunch of java processes, and on Solaris the
> microstate (/proc/pid/usage) data it collates is far more accurate and
> complete than the data you get from ps/top etc (/proc/pid/pstatus)
> 
> I'm not sure what happened to the orca config file that plotted the data,
> but workollator.se is part of the SE toolkit. It runs as a separate
> process from orcallator.se since the input data, frequency of update and
> output file format are different. There is an extra processing step called
> "explode" that takes the combined output file and generates the per-
> workload files for orca to display.
> 
> I wrote this up somewhere, but its been long enough that I can't remember
> all the details.
> Also, I did some tweaks to improve it and can't remember if they made it
> into the latest SE release.
> If there is enough interest in getting this sorted out, I'll try and dig
> out the code and get it into the orca build. Also I think Justin wrote up
> some of this in a Sysadmin mag online article.
> 
> Cheers Adrian
> 
> -----Original Message-----
> From:	orca-users-bounces+acockcroft=ebay.com at orcaware.com on behalf of
> Blair Zajac
> Sent:	Mon 8/29/2005 8:02 PM
> To:	Nguyen, Binh T
> Cc:	orca-users at orcaware.com
> Subject:	Re: [Orca-users] Snapshot 480
> 
> Nguyen, Binh T wrote:
> >
> > Allan
> > In May 2005 orca mailing list there is some discussion of monitoring
> > certain process like java http and oracle etc..
> > I check the 0.27 se file and cant find any indication of it.. I took a
> > latest snapshot and browse through the se
> > And saw a few new additions like work_load stuff so I assume it's only
> > available in the later snapshot.
> 
> Oh yes, the work_load features.  Haven't had a chance to look at that.
> Not
> having a Solaris system handy makes it tough :)
> 
> Regards,
> Blair
> _______________________________________________
> Orca-users mailing list
> Orca-users at orcaware.com
> http://www.orcaware.com/mailman/listinfo/orca-users
> 
> 
> 
> 
> _______________________________________________
> Orca-users mailing list
> Orca-users at orcaware.com
> http://www.orcaware.com/mailman/listinfo/orca-users








More information about the Orca-users mailing list