fields in sar,vmstat,top

[tempsample]

hi

for checking cpu utilization i am using sar -M 1 5

for memory utilization - vmstat 1 5

for disk utilization sar -d 1 5

i need to know what are the important fileds that I should check in vmstat and sar output.

[likid0]

Hi,

without going into detail, just to get you started.

First of all do a man sar and understand what info you get in the columns.

Sar -M shows info per cpu/core on the server. You would wan't to look if you have cpus with %idle time, if you have very little idle or 0, you should check if the server is queing up processes in the run queue with sar -q for example. if the run qsize has a high number or it's getting bigger, it's a sign the cpu is saturated.

I don't like vmstat on hpux, but as before I would say man vmstat, and read what al the columns are, and try to understand them. Quick and dirty in vmstat, you should look at the so,po and sr, any number in so and po is a sign of memory preassure, also increasing numbers in sr it also means memory preassure, and the vm working hard to free memory pages.

With sar -d, you can check the %busy column, it never should get to 80%, and the avwait(average wait time) of the io request in the queue, and avserv(average service time) to que the io request serviced, these numbers depend on the kind of IO,storage,etc.., get a baseline when all is ok, and you can have it for reference when things go awol.

this is just a little intro,not very usefull, take a look at:

http://h21007.www2.hp.com/portal/download/files/unprot/devresource/Docs/TechPapers/UXPerfCookBook.pdf