vmstat

[Tonatiuh]

HP-UX 11.23.

vmstat shows in the column r (runque waits) a number mayor than my number of CPUs. But the percentage of usage of the CPUs es 50%.

How can this happend?

What is the meaning of this situacion?

[G. Vrijhoeven]

Hi,

this could indicate an disk bottleneck.
You could check sar -uM for info on how the CPUs are used.

Regards,

Gideon

[RAC_1]

What does following give??

uptime
glance -a
sar -q 5 5

Anil

[Mark Greene_1]

In addition to what Gideon suggested, I'd also run several iterations of iostat to determine if any particular disk is being hit hard.

Do you have glance? There is, if I recall correctly, a 30 day trial on the CD of software extras that comes with the OS CD set. You can watch CPU and disk traffic in real time with that.

mark

[Tonatiuh]

$ uptime
8:50pm up 13 days, 10:59, 1 user, load average: 0.10, 0.14, 0.22

$ glance -a
/usr/bin/sh: glance: not found.

$ sar -q 5 5

HP-UX oficial1 B.11.23 U ia64 10/01/04

20:52:29 runq-sz %runocc swpq-sz %swpocc
20:52:34 2.0 10 0.0 0
20:52:39 0.0 0 0.0 0
20:52:44 1.5 20 0.0 0
20:52:49 1.0 30 0.0 0
20:52:54 2.0 10 0.0 0

Average 1.4 14 0.0 0

[RAC_1]

iostat 3 10
sar -d 5 10

Anil

[Sridhar Bhaskarla]

Hi,

I wouldn't be too concerned about the queues reported by vmstat, sar or uptime if the CPU utilization is not near 90%. It's merely an average of the samples and may not necessarily a real figure.

I would have looked for patches if the OS is of earlier versions like 10.20 etc.,. Unless your application is really impacted, I wouldn't worry much about it.

-Sri