Need help with Glance Scripting

[Dave Jadischke]

I need help stopping this glance command:

glance -advisor_only -syntax gpm_advisor -j2 >> glance.out

it works great but I need it to stop its self after 20 or so cycles, any suggestions?

[Mladen Despic · Accepted solution]

Dave,

Try this:

glance -advisor_only -syntax gpm_advisor -j2 >> glance.out &
sleep 40
kill %%

NOTE: if you run this from a script, you will likely get error messages - like "GlancePlus Non-Fatal Error" and "Unable to perform functions on devices" - but HP says this is ok.

HTH

Mladen

[Juan González]

Hi Dave,
try this
glance -advisor_only -syntax gpm_advisor
-j2 -iterations 20>> glance.out

Best regards
Juan Gonzalez

[Dave Jadischke]

Thanks for the pointers both ways work well. BUT... I'm pulling 3 different metrics, if I run it for 2 interations I get 434 lines of data in 3 columns. Where does 434 come from? I'm trying to strike an average for each metric but I?m not sure if I should add them up then divide by 434 or if there is something weird with that set of numbers. Any help would be appreciated. Thanks.

[Juan González]

Hi,
are you using a LOOP command? If you do this you get the metrics for all . If you attach your script I will be more specific?

Best regards

[Dave Jadischke]

Yes, I am using a loop in the syntax file. Here is my glance command followed by the syntax file (gpm_adviser). I'm not sure what to use in place of PROCESS LOOP, is this why i'm getting so much data?

glance -adviser_only -syntax gpm_adviser -j 5 -iterations 2 > glance.out

gpm_adviser:

PROCESS LOOP PRINT proc_cpu_total_util, proc_mem_res, proc_io_byte_rate

Thanks again.

[Dave Jadischke]

Here is the output file.

[Mladen Despic]

Dave,

Because you are using "process loop", glance will loop through all current processes on your system and output the metrics that you specify for each individual process. That would be one "iteration." You probably had 217 processes running on your system when glance was executed (217+217=434 lines of output after 2 iterations)

What are you trying to accomplish?
Can you just use metrics like GBL_CPU_TOTAL_UTIL and GBL_MEM_UTIL ?

[Dave Jadischke]

What can I use instead of PROCESS LOOP?

[Mladen Despic]

Dave,

What do you want your script to do?

If your goal is simple, e.g., display total cpu utilization on your system every minute, you will probably find GBL_xxx_xxx metrics sufficient. But if, for example, you would like to log cpu utilization by top two processes every 30 seconds, then you can modify your process loop to display just what you want.

Mladen