Monitoring -tool

[deepuanil]

Hi experts,

servers - HP PA-RISC with HP-ux 11.11

i wants to set up a monitoring and alarming tool on my servers for

cpu utilization
memory utilization
disk utilization
and
network load

monitoring intervels 1 hour

wt will be the most sutable tool for this work and how can it configure

Thanks in advance
DAmu

[Arunvijai_4]

Hello,

You can use tools like GlancePlus, check this out for more information

http://forums1.itrc.hp.com/service/forums/questionanswer.do?threadId=770017

Also, SEP's system monitoring will also be helpful to you,

http://www.hpux.ws/buildmail.hpux.text

-Arun

[deepuanil]

Hi Arun


first of all thanks for your immidiate responce


The links given are not connecting..

In glance how can i set up aalarms on excedding a threshold value for the mentioned parameters

Thanks and regards
Damu

[Vipulinux]

Hi Damu

In all unixes the most used tool for monitoring is SAR.

have a look :

http://open.itworld.com/5040/nls_unixturningsar050907/page_1.html

Cheers
Vipul

[Jov]

Hi DAmu,

It depends on what monitoring are you after? If you want to do some short term monitoring to troubleshoot an issue, then sar, vmstat and other Unix standard tools will help.

If you want a good tool from HP that which provides more details the GlancePlus (require licenses).

Now, if you're talking about long term monitoring such as trending and for reporting, then you'll need something like HP OpenView Performance Agent. Then you'll need something to present the data like OV Performance Insight.

Let us know what your requires are.


Jov

[Con O'Kelly]

Hello

If you have (OVPA) MWA installed then the you need to configure the /var/opt/perf/alarmdef to set alarms for all thresholds (CPU, Memory, disk etc).

Alarms can be sent to OVO, NNM or if you do not have these products you'll need to define an 'EXEC' action in the alarmdef file.

This 'EXEC' action can used to execute any unix command or script. This allows you to log all alarms to a defined logfile or send an email etc for all performance alarms.

The OVPA Users Guide (Chp 7) has all the information you need to configure the alarms. The default 'alarmdef' file also has plenty of examples and is well documented.
http://ovweb.external.hp.com/ovnsmdps/pdf/ovpa_users_ux.pdf

After updating the file re-start the alarm generator as follows:
# mwa restart alarm

Ensure the alarm generator is configured execute the local commands:
# agsysdb -l
(Look for Exec Actions : always or on)
see man agsysdb for more details.

Cheers
Con

[Arturo Galbiati · Accepted solution]

Hi,
I use the attached script to send by email an alert when at lest 2 resources are over the threshold. Feel fre to adapt it to your convenience.
It uses glance in adviser-only mode.

HTH,
ARt

[Arturo Galbiati]

I'm sorry ...., this is teh script

[Senthil Prabu.S_1]

Hi,
Attached a simple sysinfo script.Hope this will fulfill all your needs.


HTH,
Prabu.S

[Yogeeraj_1]

hi damu,

also consider measureware for analysis of past performances.

very useful

kind regards
yogeeraj

[Carlos Roberto Schimidt]

Hi,

Glance is better for ponctual analysis, and you can after install glance customize your applications in file /var/opt/perf/parm.

If you monitor perfomance with sendmail and your server are busy, mail will be receive with delay, not so good.

If you monitor with snmp, it´s possible save hystorical data. see http://net-snmp.org

[Jileesh Augustine]

Hi Damu,

Glance plus is the best tool for this...

I'll give you the documentation if you need...

call me...

09884527478

Regds
Jileesh

[deepuanil]

i used OPVM