disk monotoring script

[Donald C Nelson]

Does any one have a process that will alert me if a disk fails or filesystems become stale. We have clients who do not want to spend any money on monotoring packages so I have to come up with a cheap way to perform this process.

[Mridul Shrivastava]

If there is any I/O error on any disk it will be logged in syslog.log/OLDsyslog.log.
If you have EMS configured there will be an entry in event.log also.

[Donald C Nelson]

A disk may failed during the night and if no one checks on the server which is located in a data center a mile apart, no one would know if it is mirrored.

[Geoff Wild · Accepted solution]

Here's a script for stales:

# cat chkstaledsks
#!/bin/sh
# Identifies and displays stale extends on all system volume groups.
# Useful to determine defect harddisks in mirrored enviroments
# Geoff Wild

vgdisplay -v | awk '/LV Name/ { print $3 }' | xargs lvdisplay -v | grep -i -e "lv name" -e "lv status" -e stale -e '?'


Just run from cron...



Rgds...Geoff

[Torsten.]

In case of failed hardware or errors on hardware the diagnostic will take notice and will put a message in the syslog file and send a mail to root.
If you have a support contract, it is a good enhancement to install HPs ISEE software on top. This will inform you and even HP about the problem. more information:
http://h20219.www2.hp.com/services/cache/10709-0-0-225-121.html?jumpid=reg_R1002_USEN

[James R. Ferguson]

Hi Donald:

Your cheapest, free monitoring comes with your standard operating system in the form of EMS, a part of the Support Tools Bundle (STM):

http://h20293.www2.hp.com/portal/swdepot/displayProductInfo.do?productNumber=B6191AAE

You can read more about this here:

http://docs.hp.com/en/diag.html

Regards!

...JRF...

[Donald C Nelson]

Thanks everyone for your assistance