We’re monitoring nodes and 2 x switches with LibreNMS. Everything is properly being graphed and monitored but there’s one Dell Force10 s60 switch whose graph has broken parts and poller time showing that its taking avg 7000 seconds to complete the poll against this switch.
Now if we monitor this same switch with cacti, it works perfectly without any glitches in the graphs so there’s no network or server issue in this case what we think is that maybe LibreNMS is polling too much of Dell Force10 S60 switch and causing high polling time.
Can anyone please help on this matter ?
Thanks, i’ve increased these values, will monitor it for few hours and get back here with the update.
You can test your changes running manually (always as librenms user)
./poller.php -h host
I am seeing following command taking around 2200 seconds to complete :-
/usr/bin/snmpget -v2c -c ‘string’ -Oqvn -M /opt/librenms/mibs:/opt/librenms/mibs/f10 -t 9000 -r 2 udp:ip:161 .184.108.40.206.220.127.116.11.0
But if i snmpwalk manually with following command, it works well:-
snmpwalk -Os -c string -t 300 -v 2c 1 ip
And if you manually run the command that takes 2200 seconds?
We’re able to fix it. For anyone facing same problem, please upgrade your switch firmware to eradicated this problem.