LACP / Memory Leak

[lweidig]

We have two Netonix switches in our network that are both behaving the same and both "triggered" by the same event. The switches are running 1.5.9rc1 on them and we have a two port LACP link setup to a Mikrotik CCR router. It had been running perfectly fine for month's with no issues. It is running LACP-A / Fast / 32768 with key 100. The Mikrotik configuration for the bonding interface is:

Code: Select all

 /interface bonding
add comment=LAN lacp-rate=1sec mode=802.3ad name=bonding1 slaves=ether7,ether8 transmit-hash-policy=layer-2-and-3

We then added a second bonding interface to the Mikrotik (running 6.48.5 btw) and as soon as we did that started getting these errors continually in the Netoxix every 30 seconds:

Code: Select all

 LACP: LACP changed state to Active on port 1 (key 100)

The VERY odd thing is that the Netonix, other than indirectly as it is connected to the Mikrotik is not involved in that second interface. Then about once per week the memory usage decides to spike from 38-45% -> 97 - 100% and the switch reboots!

We have tried rebooting the switch on its own not under memory issues, rebooting the Mikrotik on its own, disabling LAG on ports, disabling ports,... Pretty much anything we try does not resolve the issue.

We do have other Mikrotik routers with multiple bonding interfaces that are working without issue so do not believe that is the problem, though of course it plays a part in the interactions going on.

Thoughts on how to resolve or even to troubleshoot further? Thanks!