Solaris prtdiag

Finally I got the time to implement it,
Thanks Vernon, it works like a charm.

- Roland

I changed it for XYMON only, if someone is interested...
That is changing BBHOME to XYMONHOME etc etc.

#!/usr/bin/ksh

#
# PURPOSE
# This is an very simple script/test, but extremely useful.
# It detects ANY hardware failure in Sun SPARC systems by using the standard 'prtdiag' command. Prtdiag exits with a value of
# '1' if there is an hardware error, otherwise with an exit code of '0'.
# This way, no model/platform specific (i.e. V240/V890/15K, etc.) customizations are required, as the output of 'prtdiag'
# differs on most systems. It works fine on Fujitsu-Siemens systems too.
#
# Provided by: Wim Olivier, Senior Solaris/VERITAS Engineer, AL Indigo, Johannesburg, South Africa
# sunhw

# INSTALLATION
# 1. Copy the script to $XYMONHOME/ext/xymon-prtdiag.ksh
# 2. Add it to the $XYMONHOME/etc/clientlaunch.cfg file
#

TEMPFILE=/$XYMONTMP/prtdiag.OUTPUT.$$
TEST=prtdiag
COLOR="green"

if [ "$XYMONHOME" = "" ]
then
echo "XYMONHOME is not set... exiting"
exit 1
fi

if [ ! "$XYMONTMP" ] # GET DEFINITIONS IF NEEDED, should never happen...
then
# echo "*** LOADING XYMON SETTINGS ***"
. $XYMONHOME/etc/xymonclient.cfg # INCLUDE STANDARD DEFINITIONS
fi

# What is this doing?
PANIC="1" # GO RED AND PAGE AT THIS LEVEL

PLATFORM=`uname -i`
/usr/platform/$PLATFORM/sbin/prtdiag -v > $TEMPFILE
RESULT=$?
#echo $RESULT
#
# DETERMINE RED/YELLOW/GREEN
#
if [ "$RESULT" -ne 0 ]
then
COLOR="red"
fi

#
# AT THIS POINT WE HAVE OUR RESULTS. NOW WE HAVE TO SEND IT TO
# THE XYMSRV TO BE DISPLAYED...
#

MACHINE=`uname -n`

#
# THE FIRST LINE IS STATUS INFORMATION... STRUCTURE IMPORANT!
# THE REST IS FREE-FORM - WHATEVER YOU'D LIKE TO SEND...
#
LINE="status $MACHINE.$TEST $COLOR `date`
`cat $TEMPFILE`"

$RM -f $TEMPFILE

# NOW USE THE XYMON COMMAND TO SEND THE DATA ACROSS
$XYMON $XYMSRV "$LINE" # SEND IT TO XYMONSRV

On 11/10/11 11:09 PM, Vernon Everett wrote:

Hi Roland

This is what I am using. I made a mistake though, I did not write it. I must have got confused with another script I wrote. (I confuse easily)

Credit where due though, and credit goes to Wim Olivier (See comments)

It also looks like it does pretty much the same as Gautier's script.

Only issue you may have, there is a bug in the picl daemon, and it sometime crashes. (Recent patching may have fixed this)

When that happens, all bets are off, and results could vary.

Restarting picl service normally fixes everything.

Might want to add /system/picl to the list of monitored services.

There are two good SMF monitoring scripts on the Xymonton page.

http://www.xymonton.org/monitors:smf.sh

and

http://www.xymonton.org/monitors:smf2.ksh

I prefer the 2nd one, obviously :-) but the first one is simpler to configure.

The choice is yours.

Regards

Vernon

#!/usr/bin/ksh

#

# PURPOSE

# This is an very simple script/test, but extremely useful.

# It detects ANY hardware failure in Sun SPARC systems by using the standard 'prtdiag' command. Prtdiag exits with a value of

# '1' if there is an hardware error, otherwise with an exit code of '0'.

# This way, no model/platform specific (i.e. V240/V890/15K, etc.) customizations are required, as the output of 'prtdiag'

# differs on most systems. It works fine on Fujitsu-Siemens systems too.

#

# Provided by: Wim Olivier, Senior Solaris/VERITAS Engineer, AL Indigo, Johannesburg, South Africa

# sunhw

# INSTALLATION

# 1. Copy the script to $BBHOME/ext/sunhw.sh

# 2. Add it to the $BBHOME/etc/bb-bbexttab file

# 3. Restart the BB client

#

BBPROG=sunhw; export BBPROG

TEMPFILE=/$BBTMP/sunwh.OUTPUT.$$

TEST="sunhw"

COLOR="green"

if [ "$BBHOME" = "" ]

then

echo "BBHOME is not set... exiting"

exit 1

fi

if [ ! "$BBTMP" ] # GET DEFINITIONS IF NEEDED

then

# echo "*** LOADING BBDEF ***"

. $BBHOME/etc/bbdef.sh # INCLUDE STANDARD DEFINITIONS

fi

PANIC="1" # GO RED AND PAGE AT THIS LEVEL

PLATFORM=`uname -i`

/usr/platform/$PLATFORM/sbin/prtdiag -v > $TEMPFILE

RESULT=$?

#echo $RESULT

#

# DETERMINE RED/YELLOW/GREEN

#

if [ "$RESULT" -ne 0 ]

then

COLOR="red"

fi

#

# AT THIS POINT WE HAVE OUR RESULTS. NOW WE HAVE TO SEND IT TO

# THE BBDISPLAY TO BE DISPLAYED...

#

MACHINE=`uname -n`

#

# THE FIRST LINE IS STATUS INFORMATION... STRUCTURE IMPORANT!

# THE REST IS FREE-FORM - WHATEVER YOU'D LIKE TO SEND...

#

LINE="status $MACHINE.$TEST $COLOR `date`

`cat $TEMPFILE`"

$RM -f $TEMPFILE

# NOW USE THE BB COMMAND TO SEND THE DATA ACROSS

$BB $BBDISP "$LINE" # SEND IT TO BBDISPLAY

On 11 October 2011 10:12, Roland Soderstrom <user-0cec9512a49f@xymon.invalid> wrote:

Hi,

I have used old good bb-prtdiag for many years.
It works fairly well for a lot of old Sun machines but lack the new ones.
The script is pretty nasty in nested loops and too many if's.
It is not easy to update it to accommodate new machines.

Is there anyone out there that has a new version of it?
Can't find any on xymonton.
Or do you monitor SUN hw in another way than prtdiag -> xymonserver?

Otherwise I will rewrite it to support our new hw.

- Roland

Xymon@xymon.com