Average memory utilization

[Vishu]

Hi,

please help in taking the average memory utilization of HP-UX 11.00 server for one week. As for disk and CPU usage, we can have from sar command. which command we will use for memory.

Thanks

[Sharma Sanjeev]

Hi vishu

The commands sar and vmstat

For example, sample CPU utilisation every 2 seconds for 5 iterations:

sar -u 2 5

or For memory utilisation every 2 seconds for 5 iterations:

vmstat 2 5

look at the avm and free fields

Regards
Sanjeev

[Michal Kapalka (mikap)]

hi,

use glance or gpm ( gui )

mikap

[Vishu]

Thanks Sanjeev and Michal,

But i dont have the glance installed in my server...

and if vmstat helps, then i need it for one complete week. By anyway, i can add it to cron or any other way.

Thanks

[Michal Kapalka (mikap)]

hi,

i don't like top but in some cases could be used.

maybe the top will help you.

mikap

[Steven E. Protter]

Shalom,

the sar commands will work.

This script will give you a complete performance snapshot using sar that will meet your needs:

http://www.hpux.ws/?p=6

No cost uses sar

production code.

Glance is available on your OS CD for 60 day evaluation.

SEP

[Vishu]

in top output, i found following line as memory


Memory: 772800K (271340K) real, 585612K (233832K) virtual, 669856K free Page# 1/15

can you please tell me in detail, what each field is showing. My server RAM is 4096MB, but i have not seen anything in GB in this output.

Plz suggest.

[James R. Ferguson]

Hi Vishu:

On of the "best" tools for doing what you want to do is 'glance' and 'measureware'. The reason is that 'glance' is architected to "understand" the HP-UX kernel. 'top' and 'sar' are more generic UNIX tools. With 'glance'/'measureware' you can collect and graph long term metrics,

Regards!

...JRF...

[Vishu]

Thanks James,

but i am trying with glance too on one server and trying to extract the data for today, but it is replying with the error :-

# ./extract -xp -r /tmp/moni -g -b today 14:00 -e today 15:00 -f stdout


Unable to read record from logfile, size returned = 0
Unable to read record from logfile, size returned = 0
Memory fault(coredump)

Plz suggest.

[Nido]

Hello Vishu,

>>>>Unable to read record from logfile, size returned = 0
>>>>Memory fault(coredump)

Check the status of measureware daemons:
#mwa status

Check the log files in:
/var/opt/perf/datafiles/log*

if its showing zero in size; try restarting mwa:

#mwa stop ; mwa start

Let me know the outcome.

Cheers!!

[Vishu]

Well Nido,

Here are the outputs of the commands :-

root # mwa status
MeasureWare scope status:
WARNING: scopeux is not active (MWA data collector)

MeasureWare background daemon status:
(Should always be running when the system is up)
Running ttd (Transaction Tracker daemon) pid 1372

MeasureWare server status:
WARNING: alarmgen is not active (alarm generator)
WARNING: agdbserver is not active (alarm database server)
Running perflbd (location broker) pid 28562

The following data sources have running repository servers:
PID DATA SOURCE
Running rep_server 29803 SCOPE
root #


root # cd /var/opt/perf/datafiles
root # ll log*
-rw-r--r-- 1 root root 3348296 Jul 1 18:55 logappl
-rw-r--r-- 1 root root 10431046 Jul 1 18:55 logdev
-rw-r--r-- 1 root root 1391798 Jul 1 18:55 logglob
-rw-r--r-- 1 root root 3192 Jul 1 18:55 logindx
-rw-r--r-- 1 root root 20882164 Jul 1 18:55 logproc
-rw-r--r-- 1 root root 3954348 Jul 1 18:55 logtran
root # date
Wed Jul 1 19:18:37 CDT 2009
root #

root # ps -ef | grep -i mwa
root 4611 4284 0 19:19:33 pts/tb 0:00 grep -i mwa
root #


One query, i grep for mwa process, but did not find any. then how come the log files are updated till 1st July. Plz put some light on it.

thanks

[Nido]

Hi Vishu,

>>>>WARNING: scopeux is not active (MWA data collector)
>>>>WARNING: alarmgen is not active (alarm generator)
>>>>WARNING: agdbserver is not active (alarm database server)

Try restarting mwa:
mwa stop ; mwa start

If its not solving the problem:
Then first copy all log files to different location and empty the logs.

ll /var/opt/perf/datafiles/log*

Try restarting mwa again:

Let me know hows the extract output.

Good Luck!!

[Karl oliver]

hi Vishu
I use a utility
memdetail . I think i got the utility from this forum. sample output below
You could put this in crontab .
The output from this is the same as glance provides
memdetail

Memory Stat total used avail %used
physical 16357.0 11153.1 5203.9 68%
active virtual 5279.8 4959.4 320.4 94%
active real 4438.3 4203.1 235.1 95%
memory swap 16357.0 6995.2 9361.8 43%
device swap 16384.0 4662.4 11721.6 28%

[Vishu]

Karl,

Can you please provide the full path to this memdetail utility. and the syntax to use it.

Thanks

[Karl oliver]

sorry I dont know location of code . But someone on this forum will.
Maybe you could open a thred asking this ?
Karl

[Vishu]

Can anyone help in getting this memdetail utility and usage syntax, which Karl is telling about.

Thanks

[Dennis Handly · Accepted solution]

>Can anyone help in getting this memdetail?

Using a forums search finds Geoff's reply:
http://forums.itrc.hp.com/service/forums/questionanswer.do?threadId=1179723

[Vishu]

Thanks Dennis...

i got what is required... Really helpful..full points.

Thanks everyone for your inputs on it.

[Vishu]

closing the thread.

Thanks