We have a MikroTik router at a colo, running RouterOS 5.9. One Ethernet interface connects to the Internet, another to an MPLS cloud which connects to our other three sites. This router was fine up until last week, when the provider migrated us and other customers to a new router on their side. There was a day-long outage while they figured out what they did wrong, then everything came up fine–but now our router needs rebooting every 24 hours or so. It always responds to pings from outside, but does not pass traffic from one side to the other for more than about 24 hours. Power-cycling it always fixes the problem–for a day.
Ideas?
I’m putting this in the Beginner Basics forum because my background is mostly Cisco, never played with these boxes before.
Thanks!
If I understood right, you need reboot the router 1 time for a day:
/system scheduler
add interval=1d name=auto_reboot on-event="/system reboot" policy=\
ftp,reboot,read,write,policy,test,winbox,password,sniff,sensitive,api \
start-date=oct/10/2014 start-time=03:00:00
Maybe disable/enable the interface resolve the problem.
Try It,if problem resolved you can write a script for disable/enable it.
Oddly, the router seems OK when I look at it: utilization is low, one interface (internal/MPLS) is reporting errors but that could be due to bandwidth limitations, and the log shows no errors.
However, the cache contains 29 entries for “www.jrdga.info”, an address associated with DNS amplication attacks and DDOS attacks.
Any ideas?
The plot thickens: the errors go away when I force the interface to 100Mbps/Full-duplex, which it should auto-negotiate but never does.
However, after the interface runs error-free for a day or so, it locks up, stops passing packets, and stays that way until I reboot the router. (Blinking, disabling/enabling the interface still does nothing.)
For now I’m implementing bingo220’s daily automatic reboot suggestion.
Thanks to all!