Health Check of the Servers

[Shivkumar]

Hi,

What are the various ways to check the health of the systems ?

Thanks,
Shiv

[freddy_21 · Accepted solution]

using diagnostic tool
# mstm



thanks
freddy

[inventsekar_1]

lot of softwares available.

my first thought is about the great Openview Operations (OVO)
http://h20229.www2.hp.com/
OVO costs much and with that you can manage a total IT environment.


and mstm informations:
STM Tutorial: mstm Commands
http://docs.hp.com/en/diag/stm/stt_mcomm.htm

http://docs.hp.com/en/30216-90325/ch07.html

[Raj D.]

Shiv,

Here is a good link with all info:

http://forums1.itrc.hp.com/service/forums/questionanswer.do?threadId=1047010


Cheers,
Raj.

[Rajesh SB]

Hi Shiv,

For Hardware health check, the best tool is cstm/mstm/stm.

For routine health check is monitoring,

Disk Usage across different file system.
# iostat

Memory Utilization and bottleneck monitoring
# vmstat -n

For system

Value added tools from HP are

# HP Glance Pak/Measureware which helps to monitor system kernel parameter usage,disk,cpu, memory/swap and processes monitoring.

OpenView Operation - IT Management software.
customize to monitor system health monitoring and log message monitoring like syslog and application logs.

Regards,
Rajesh

[Raj D.]

Shiv ,
Also dont forget to keep an eye on /var/adm/syslog/syslog.log and EMS log file /var/opt/resmon/log/event.log


Cheers,
Raj.

[Steven E. Protter]

Shalom Shiv,

Built into the OS for free of course is the EMS system.

What most system administrators do is write or modify scripts to check what is important to them.

Over my early years as an admin my systems were plagued by disk issues. It always began with an lbolt in the syslog and ended with a hug of war with support and a disk replacement.

So every night at 4 a.m. the following ran.

err=$(grep -i lbolt|wc -l)
if [ $err -ge 1 ]
then
echo "Hardware problem" |mailx -s "Hardware fault" myemail@myoldjob.org
fi

That at least gave me a head start and usually the backups were done by 4 a.m. so it was an ideal time to run the check.

You tailor the check to meet the needs of your systems, your personal life and your users, though not in that order.

SEP

[Raj D.]

Shiv,

Also you might write a script to monitor the health of the disks/lvols, and run every 5 min through cron.

The stirng "stale" can be use with the lvdisplay withing a script, and mailx the result if it finds any stale pe.

# lvdisplay -v $lvol_name | grep stale

So if any stale PE it finds, it will notify you , by email.


Cheers,
Raj.

[Robert Fritz]

Try "system health check":

http://h20293.www2.hp.com/portal/swdepot/displayProductInfo.do?productNumber=SHCBASE01

It covers a wide range.

-Robert