vmstat

[Waugh]

Dear All,

What i found that the vmstat and few other commands taking more time to excute it. What will be the reason .

time vmstat 1 3
kthr memory page disk faults cpu
r b w swap free re mf pi po fr de sr m1 m1 m1 m2 in sy cs us sy id
0 3 0 42505288 20739608 424 523 185 19 19 0 0 4 3 3 0 6245 35933 4939 5 1 94
0 11 0 41885320 19187776 1829 843 0 49 49 0 0 1 1 1 0 36877 193330 66223 17 5 78
0 12 0 41892712 19187768 55 283 0 81 81 0 0 9 9 9 0 35711 90055 63995 14 4 82

real 0m31.38s
user 0m27.61s
sys 0m1.74s

Regarda,
Waugh

[Dennis Handly]

Taking more time than when? What HP-UX version?

[Michael Steele_2]

a) UNIX95=1 ps -ef -o pid,ppid,wchan,stat | grep PID
wchan checks to see what is blocking the pid and what it is waiting on

b) If you review most hpux admins notes on vmstat they will tell you it is mostly unreliable on hpux since 10.00 when the transition from BSD to ATT was made. Most rely on glance / gpm, and top.

[Dennis Handly]

>Michael: UNIX95=1 ps -ef -o pid,ppid,wchan,stat | grep PID

If you want a particular PID you may want to use -p:
UNIX95=EXTENDED_PS ps -f -p your-PID -o pid,ppid,wchan,stat

[Yogesh M Puranik]

Hello Waugh,

As you mentioned that few commands like vmstat taking more time to run.Below could be the reasons which of that cause.

As your vmstat output clearly concludes that page outs are more than page in, so there could be the issue of memory crunch.

1]Please check whther the physical memeory & swap utilisation is 100% full.
#vmstat or swinfo -tam

2]Also check if /tmp file system is 100% full.In this scinario as well there could be the possibilities of commands getting hung or poor response.
#bdf /tmp

3]One more parameter you can check is load average of the system.if it is more than avg 3.0 then it implies system is loaded.
#top or sar 1 5

Hope this will help you.

Regards
Yogesh

[andrewstolpe]

I'm seeing the same results as Waugh. My physical memory and swap utilization look good and I followed your other checks and still no difference. What do you think it could be?

[Dennis Handly]

>I'm seeing the same results as Waugh. My physical memory and swap utilization look good

You might want to explain your results?  Are you worried about your system performance or about the slowness of the vmstat command?