Not sure if this has been covered before....
I am pinging a bunch of servers over a VPN connection and have been having problems using icmpping. There are many times that the latency on the VPN will go way up and fping times out before the response is received causing a erroneous failure alert.
Evidently the default timeout in fping is 500ms.
You can change this by editing the source file
~/src/zabbix_server/pinger/pinger.c
and change
zbx_snprintf(str,sizeof(str),"cat /tmp/zabbix_server_%li.pinger | %s -e 2>/dev/null",
to something like
zbx_snprintf(str,sizeof(str),"cat /tmp/zabbix_server_%li.pinger | %s -t 5000 -e 2>/dev/null",
'-t 5000' increased the timeout to 5 seconds and completely eliminated the bogus alerts I was getting.
Maybe this should be added as a configurable parameter in Zabbix?
Duane
I am pinging a bunch of servers over a VPN connection and have been having problems using icmpping. There are many times that the latency on the VPN will go way up and fping times out before the response is received causing a erroneous failure alert.
Evidently the default timeout in fping is 500ms.
You can change this by editing the source file
~/src/zabbix_server/pinger/pinger.c
and change
zbx_snprintf(str,sizeof(str),"cat /tmp/zabbix_server_%li.pinger | %s -e 2>/dev/null",
to something like
zbx_snprintf(str,sizeof(str),"cat /tmp/zabbix_server_%li.pinger | %s -t 5000 -e 2>/dev/null",
'-t 5000' increased the timeout to 5 seconds and completely eliminated the bogus alerts I was getting.
Maybe this should be added as a configurable parameter in Zabbix?
Duane
Comment