vmstat: negative values?!

[Alberto Rocha]

Hello, i'm monitoring a largely stressed system. That's an rp5470 (L3000-5xx) w/ 2 550MHz CPU 4 GB of mem running a J2EE app and an Oracle 9i database.

What is confusing me now is the negative values in the response to the "vmstat -f" command :

# vmstat -f
3031208 forks, -433171143 pages, average= -142.90

Any hints?
Thx in advance

[Orhan Biyiklioglu]

The values in vmstat are signed data types so they can overflow as the counter gets larger. This should appear if your system is up for a very long time or the memory page swapping is too high.

hth

[Arunvijai_4]

http://unix.derkeiler.com/Mailing-Lists/HP-UX-Admin/2005-06/0017.html

It says, ..."vmstat tries to present an average since boot on the first line. The
first line of vmstat output is utter garbage and must be discarded"

I was only running vmstat for a single count. When run with mutiple counts
the stats after the first line are reported accurately. ""

-Arun

[Borislav Perkov]

Hi Alberto,

It is overflow.
Try first to clear with

vmstat -z

Regrads,
Borislav

[Alberto Rocha]

Thank you for all the replies. In fact i suspected of some kind of overflow but now i have the confirmation.
AR