HP Insight Control Software for Linux - Manual (Page 226)

Corrective Actions Cause/Symptom
The Nagios default threshold values for total
processes
, user processes, and zombie
nagios_vars.ini file. HP recommends saving a copy
of the original file before making any updates.
processes might be too small for certain system 1. Save a copy of the original file:
configurations, particularly those with virtualization
# cp /opt/hptc/nagios/etc/nagios_vars.ini
/opt/hptc/nagios/etc/nagios_vars.ini.orig
operating systems. If so, you will encounter CRITICAL or
WARNING
alerts for the NodeInfo service.
2. Edit the nagios_vars.ini file to update the values
for the *_procs_warning and *_procs_critical
parameters to values that apply to your environment.
For example:
[default]
total_procs_warning = 400
total_procs_critical = 500
user_procs_warning = 200
user_procs_critical = 300
zombie_procs_warning = 1
zombie_procs_critical = 5
[service_nodes]
total_procs_warning = 500
total_procs_critical = 600
user_procs_warning = 300
user_procs_critical = 400
zombie_procs_warning = 1
zombie_procs_critical = 5
3. Rebuild the /opt/hptc/etc/sysconfig/vars.ini
file and push out the new vars.ini file to all the
managed systems:
#
/opt/hptc/nagios/libexec/check_nagios_vars
--rebuild
#
/opt/hptc/nagios/libexec/check_nagios_vars
--update
Vars Warning - iclx[2-6] vars.ini have
been resynchronized
NOTE: The vars.ini file is dynamically created from
the nagios_vars.ini file, so any changes to the default
Nagios thresholds must be made to nagios_vars.ini,
not to vars.ini.
See “Resolving host names on the CMS” (page 78) Nagios host check shows host status as "DOWN"
Nagios incorrectly reports a managed system as down
when the host name for an existing server in HP SIM
database has changed since Options
→IC-Linux→Configure
Management Services was last run.
Edit the /opt/hptc/nagios/etc/syslogAlertRules
file on the CMS and update the mcelog_error rule as mcelog Nagios alerts for the Syslog Alerts service on
SLES11 SP1 managed systems
Nagios periodically calls the /usr/sbin/mcelog utility
on managed systems. There is s a known issue in which
shown below so that Nagios alerts are not generated for
the mcelog: Cannot mmap SMBIOS tables syslog
event.
programs using mmap to read SMBIOS info (like mcelog)
fail when hp-health is installed.
rule mcelog_error {
name (!/Cannot mmap SMBIOS tables at/)
For example, the following mcelog messages appear in
the /var/log/messages and will cause Nagios to
generate an alert.
relevance ($subsystem =~ /mcelog/)
226 Troubleshooting