Robert G. Brown rgb at
Thu Aug 28 16:42:19 EDT 2003

On 28 Aug 2003, Joseph Landman wrote:

> Hi Glen:
>   Several methods.
> 1) vmstat
> 	vmstat 1
> and look at the so/si columns, not to mention the r/b/w.
> 2) swapon -s
> to see the swap usage
> 3) top
> has an ok summary of the vm info
> 4) cat /proc/meminfo 
> can give a crude picture of the memory system.

and if you want to watch pretty much all of this information in parallel
(on all the systems at once) xmlsysd provides output fields with the
information available in both vmstat and free (cat /proc/meminfo), so
you can actually watch for swapping or paging or leaks on lots of
systems at once in wulfstat.  It easily handles updates with a 5 second
granularity and can often manage 1 second (depending on your network and
number of nodes and so forth).

It's on the brahma website or linked under my own.

I don't really provide a direct monitor of disk activity (partly out of
irritation at custom-parsing the multidelimited "disk_io" field in
/proc/stat), but if you were really interested in it I could probably
bite the bullet and add a "disk" display that would work for up to four
disks in a few hours of work.

I'd guess that ganglia could also manage this sort of monitoring as
well, but I don't use it (as I wrote my package before they started
theirs by a year or three) so I don't know for sure.


Robert G. Brown	             
Duke University Dept. of Physics, Box 90305
Durham, N.C. 27708-0305
Phone: 1-919-660-2567  Fax: 919-660-2525     email:rgb at

Beowulf mailing list, Beowulf at
To change your subscription (digest mode or unsubscribe) visit

More information about the Beowulf mailing list