Unable to communicate with alarm infrastructure

book

Article ID: 167981

calendar_today

Updated On:

Products

XOS

Issue/Introduction

This article describes the processes involved in the alarm infrastructure on the CPM and possible causes for the message "Unable to communicate with alarm infrastructure".

Commands accessing the alarm infrastructure report following message:
Unable to communicate with alarm infrastructure at this time...

This message is shown directly in the CLI using the following command:
show alarms active

At the same time, the messages file may contain the following message:
cbshmonitord: [E] Unable to get alarm stats from alarm manager

 

Cause

The above messages may be caused by one of daemons not running or working properly. It may also be related to the situation documented in Knowledge Base article 6044 "WARNING: database "running-config_alt" must be vacuumed". If the latter is the cause, the database must be vacuumed first.

Resolution


Workaround

The following processes are used for alarm reporting:

cbsalarmmgrd
cbsactivemqd
cbshshimd


Status of these processes is monitored by the cbsinitd process. If the cbsinitd process discovers that one of monitored processes is not running, it will kill all remaining running pids for that process and start a new one. This is recorded in the messages file and reported in the following entries:


Aug  2 06:51:02 conwell cbsinitd[4516]: [W] [conwell 1.1.6.20] daemon 'daemon.cbsalarmmgrd' pid '1106' not running (3:No such process)
Aug  2 06:51:02 conwell cbsinitd[4516]: [I] [conwell 1.1.6.20] killing any remaining running pids for daemon 'daemon.cbsalarmmgrd' (prcnam:cbsalarmmgrd)
Aug  2 06:51:04 conwell cbsinitd[4516]: [I] [conwell 1.1.6.20] starting daemon 'daemon.cbsalarmmgrd' using command '/etc/rc.d/init.d/cbsalarmmgrd start'
Aug  2 06:51:06 conwell cbshmonitord[4330]: [E] [conwell 1.1.6.20] Unable to get alarm stats from alarm manager
Aug  2 06:51:06 conwell cbsinitd[4516]: [I] [conwell 1.1.6.20] daemon 'daemon.cbsalarmmgrd' (prcnam:cbsalarmmgrd) is running with 1 pid(s)
Aug  2 06:51:11 conwell cbshmonitord[4330]: [E] [conwell 1.1.6.20] Unable to get alarm stats from alarm manager


Stopping any of the above mentioned processes will trigger cbsinitd to restart that process. All alarm reporting is reset, and the error message "Unable to communicate with alarm infrastructure" should not be reported anymore.