Bull Escala Power7 Series, Manual

Page: 57 / 148

If the system to which the enclosure is attached has a Hardware Management Console (HMC), or a
Systems Director Management Console (SDMC), or is running the Electronic Service Agent
™
application, create a file named run_ses_healthcheck in the
/usr/lpp/diagnostics/bin
directory, with
the following contents:
------------------------------------------------------------------
#!/bin/ksh
#Name:run_ses_healthcheck
#Location:/usr/lpp/diagnostics/bin
#Function: SCSI SES hourly heathcheck
for i in ’lsdev -Cc container -t ses -s scsi -F name -S available’
do
diag -cd $i > /dev/null
done
#any registered "external notification" will be notified of errors
#(such as HMC or SDMC or Electronic Service Agent)
------------------------------------------------------------------
Note: On a logically partitioned system, the CRON job and script only need to run on one partition
that is associated with the enclosure. However, it is acceptable to have the same CRON job and scripts
on any partition that is associated with the enclosure.
Example 2
If the system to which the enclosure is attached has no HMC, has no SDMC, and is not running the
Electronic Service Agent application, create a file named run_ses_healthcheck in the
/usr/lpp/diagnostics/bin
directory, with the following contents:
------------------------------------------------------------------------
#!/bin/ksh
#Name:run_ses_healthcheck
#Location:/usr/lpp/diagnostics/bin
#Function: SCSI SES hourly heathcheck
for i in ’lsdev -Cc container -t ses -s scsi -F name -S available’
do
diag -cd $i > /dev/null
if [$? -ne 0]
then
/usr/lpp/diagnostics/bin/diagrpt -o >/tmp/ses.health.output
#you might want to process the output prior to placing it in
#a file
#somhow notify the user of the error.
A sample is shown
#below.
mail -s "7031 Health Check" root</tmp/ses.health.output
rm /tmp/ses.health.output
fi
done
------------------------------------------------------------------------
Note: You can customize the actual notification in the script based on your preferences. If errors need
attention, these scripts send mail to the root user. You can modify the script to inform certain users
about the errors.
In both examples, you have created the run_ses_healthcheck file. To run this file, type the following
command at the AIX command prompt:
chmod 544 /usr/lpp/diagnostics/bin/run_ses_healthcheck
Connecting and configuring the 31D/24, or 31T/24 SCSI disk drive enclosure in an
AIX clustered environment
Special considerations exist for connecting the disk drive enclosure in a clustered environment. The
information in this section can help you meet these considerations.
Disk drive enclosures
45