So simple but so frustrating!!
Hardware: RB1000, firmware V4.5
The current configuration is very basic, so basic it cannot possibly be a problem but everything is pointing to routing?
Ethernet 1 - UnTagged switch connection subnet 192.168.102.0/24
Ethernet 2 - UnTagged switch connection subnet 172.17.191.0/24
Ethernet 3 - UnTagged switch connection subnet 192.168.100.0/24
Most routing is dynamic. There is a default route defined - 172.17.191.254 - and there are 2 static routes defined pointing to remote networks connected via a Cisco Router on the Ethernet 3 subnet.
Problem:
Routing to/from the Eth2 subnet stops for anything from sub second to a few (<5) seconds. That is neither Eth1 not Eth3 can contact any host on the Eth2 subnet.
A host on Eth2 subnet looses contact with hosts on Eth1 and Eth3 at exactly the same time. This host does not loose contact with hosts on its own subnet - 172.17.191.0 at any time which seems to eliminate the possibility of a physical or switch problem on the Eth2 subnet.
A host on Eth2 does not loose contact with the Mikrotik router port on its subnet, although the only anomaly I have seen thus far is something I have never seen before - when the routing is working the ping response from the Mikrotik port - 172.17.191.253 - reports a TTL=64, but as soon as the routing problem occurs this ping response succeeds but the TTL is now 255 ??
The time between routing outage does vary between 6 and 18 minutes, but is predominantly 9 minutes. Again curious.
Anyone seen anything like this before?