Global Run Queue

[OneNeck UNIXSA]

Is there a way to list processes that reside in the global run queue?

[George A Bodnar]

Not sure this is what you want, but this shows you runnable processes:

(export UNIX95=""; ps -eo pid,state,args | grep " [R] ";)

[OneNeck UNIXSA]

I can use ps -efl to see states, was looking more at which processes were in the run queue. Thanks for the input though.

[Martin Johnson]

George's way is thje only way I know how to determine who's waiting to run.

You understand that the Global run queue is a moving target. The entries are constantly changing the the processes move through the queue.

Is there something in particular that you are looking for?

HTH
Marty

[OneNeck UNIXSA]

I was looking for a way to identify the processes in the run queue when MWA makes and alert. Currently I use ps -efl and look for the running and waiting processes, but this doesn't seem to show the information that was looking for. I have mwa setup to alert such as:
Global CPU Utilization = 66.33%
Global CPU System Mode Util should be < 50 = 38.21%
Global CPU Interrupt Util should be < 30 = 0.85%
Global CPU Context Switching Util should be < 30 = 1.79%
Global Priority Queue should be < 3 = 0.24
Global Run Queue should be < 3 = 19.00
Detected on 08/15/02 at 07:00

and just wanted to get the threshold and any extended data. Just tweaking. :)

[Martin Johnson]

I alert on a combination of CPU and global run queue:

symptom CPU_Bottleneck type=CPU
rule GBL_CPU_TOTAL_UTIL > 90 prob 35
rule GBL_CPU_TOTAL_UTIL > 95 prob 25
rule GBL_CPU_TOTAL_UTIL > 98 prob 15
rule GBL_CPU_TOTAL_UTIL > 100 prob 8
rule GBL_RUN_QUEUE > 2 prob 10
rule GBL_RUN_QUEUE > 4 prob 15
rule GBL_RUN_QUEUE > 8 prob 25

alarm CPU_Bottleneck > 80 for 5 minutes
type = "CPU"
start
if CPU_Bottleneck > 90 then
red alert "CPU Bottleneck probability= ", CPU_Bottleneck, "%"
else
yellow alert "CPU Bottleneck probability= ", CPU_Bottleneck, "%"
repeat every 10 minutes
if CPU_Bottleneck > 90 then
red alert "CPU Bottleneck probability= ", CPU_Bottleneck, "%"
else
yellow alert "CPU Bottleneck probability= ", CPU_Bottleneck, "%"
end
reset alert "End of CPU Bottleneck Alert"

HTH
Marty

[OneNeck UNIXSA]

I think I am going to go with:

alarm (GBL_CPU_TOTAL_UTIL > 90) or
(GBL_CPU_SYS_MODE_UTIL > 50) and
(GBL_PRI_QUEUE > 3) or
GBL_RUN_QUEUE > 4) for 5 minutes

Still looking for a way to list the run queue though. :)

Previously my and was an "or".