We very often have the problem when restarting the Zabbix agent that the agent processes get "stuck", and Zabbix fails to start again. We need to kill the processes, and then start Zabbix. This is on Debian Lenny stable.
We switched to using "invoke-rc.d zabbix-agent stop && sleep 10 && invoke-rc.d zabbix-agent start" but it still occurs:
Any suggestions to avoid this? This happens randomly on different machines.
We switched to using "invoke-rc.d zabbix-agent stop && sleep 10 && invoke-rc.d zabbix-agent start" but it still occurs:
Code:
22990:20091002:003204 Got signal. Exiting ... 22991:20091002:003204 Got signal. Exiting ... 22995:20091002:003204 Got signal. Exiting ... 22990:20091002:003204 zbx_on_exit() called. 22994:20091002:003204 Got signal. Exiting ... 22992:20091002:003204 Got signal. Exiting ... 22993:20091002:003204 Got signal. Exiting ... 22990:20091002:003204 One child process died. Exiting ... 22990:20091002:003204 zbx_on_exit() called. 22990:20091002:003204 Cannot remove PID file [/var/run/zabbix-agent/zabbix_agentd.pid] [No such file or directory] 22990:20091002:003206 ZABBIX Agent stopped 21603:20091002:003216 zabbix_agentd started. ZABBIX 1.4.6. 21604:20091002:003216 zabbix_agentd collector started 21605:20091002:003216 zabbix_agentd listener started 21606:20091002:003216 zabbix_agentd listener started 21607:20091002:003216 zabbix_agentd listener started 21608:20091002:003216 zabbix_agentd active check started [hostname.com:10051]
Comment