RE: [Ganglia-developers] RFC: IO-Wait CPU metrics for all platfor ms

[Martin K. <kn...@kn...>]

Yemi,

 2.6 is your friend :-) If I read the source right, you find the
following numbers in /proc/stat: user, nice, system, idle, iowait, irq,
softirq

 2.4 had only: user, nice, system, "all other jiffies not in a
category"

 The numbers are of course per-cpu and not per-partition.

Cheers
Martin
--- "Adesanya, Adeyemi" <ye...@sl...> wrote:
> 
> Martin,
> 
> I have been missing this kernel info too. I wrote some of my own
> custom Ganglia disk IO metrics that are 'per-partition' just like
> iostat. I get the data from /proc/partitions but so far I have been
> unable to calculate 
> "%wait" for a partition (percentage of elapsed time wait queue was
> not zero).  I really want this on Linux so my stats are consistent
> with Solaris. 
> 
> ----
> Yemi
> 
> 
> > -----Original Message-----
> > From: Martin Knoblauch [mailto:kn...@kn...] 
> > Sent: Tuesday, May 25, 2004 1:30 PM
> > To: gan...@li...
> > Subject: [Ganglia-developers] RFC: IO-Wait CPU metrics for 
> > all platforms
> > 
> > Hi,
> > 
> >   the 2.6.x Linux kernel exports statistics on the amount of 
> > time that the CPU is waiting for I/O operations. I think it 
> > would be useful to monitor this with gmond. In the minumum 
> > case we probably have to retrieve it and add it to the idle
> numbers.
> > 
> >  From the other platforms only HP-UX and Solaris report the 
> > IO-wait cycles. Unfortunatelly under different names :-(
> > 
> >  My proposal would be to:
> > 
> > - monitor IO-wait for "all" platforms. If the platforms don't 
> > have it, just report "0"
> > - report it under the same name on all platforms
> > 
> >  My reason for the "all platform" metrics is to allow it into 
> > the webfrontend CPU report graph. We already do it for 
> > "nice", which is not exported by all platforms.
> > 
> >  I would volunteer to write the stub routines for all 
> > platforms and to modify the webfrontend.
> > 
> >  Comments? Am I out of my mind? The major drawback is that it 
> > adds some kind of incompatibility to pre-2.6 code. On the 
> > other hand 2.6 is the "right" place to do it.
> > 
> > Cheers
> > Martin
> > 
> > =====
> > ------------------------------------------------------
> > Martin Knoblauch
> > email: k n o b i AT knobisoft DOT de
> > www:   http://www.knobisoft.de
> > 
> > 
> > -------------------------------------------------------
> > This SF.Net email is sponsored by: Oracle 10g Get certified 
> > on the hottest thing ever to hit the market... Oracle 10g. 
> > Take an Oracle 10g class now, and we'll give you the exam FREE.
> > http://ads.osdn.com/?ad_id=3149&alloc_id=8166&op=click
> > _______________________________________________
> > Ganglia-developers mailing list
> > Gan...@li...
> > https://lists.sourceforge.net/lists/listinfo/ganglia-developers
> > 
> 
> 


=====
------------------------------------------------------
Martin Knoblauch
email: k n o b i AT knobisoft DOT de
www:   http://www.knobisoft.de