Our Splunk Server has a dedicated subinterface for listening on udp port 514 and I belive most customers does the same implementation.
Cause udp is only uni directional traffic, after 300 seconds the switches lost the mac-address of the splunk server while the default gateway has normaly a much longer arp aging time ( 4h on cisco devices ). So there is no need for the gateway to refresh the arp cache but the switches start flooding the traffic to every port in the vlan after the mac address has aged out.
My question is:
Does Splunk provide a mechanism out of the box to avoid this problem? Or should I create a cron job which sends gratious arps every 300 seconds?
Thanks
Splunk provides no solution for this. Just set up a cron job to ping something every so often.
Or, add your syslog box to your monitoring solution and ping it / check_ssh on it every couple of minutes. Added Bonus: you'll know when it's down.