I have 2 RB250 switches set up as ascii illustrated below:
[ RB450G ]—trunk----[ Cisco SW ]----trunk----[ RB250-1 ]—trunk—[ RB250-2 ]—[ devices ]
I have the RB450G set to netwatch the Cisco and the 2 RB250’s as well as a device connected to the last RB250.
The netwatch is every 10 seconds and 100ms timeout. Ping times are under 10ms.
RB250-2 has constant netwatch failures. It is random, but lasts for sometimes a minute or more.
I can manage the RB250’s using the web interface.
The devices connected to RB250-2 never have a netwatch failure.
None of the other switches in the chain have a netwatch down event when the RB250 does.
Any ideas?
Is there any NAT configured in this setup?
No NAT.
I manage the switches via vlan30 which is assigned to the RB450G port and has its own bridge with ip address 10.20.2.1/24.
The netwatch is set to the local lan address of the switches 10.20.2.x with 10 seconds ping interval, 100ms timeout.
Ping times are 3ms or less.
I manage the Cisco switches (4 of them) on this same vlan.
The Cisco’s have not missed a single netwatch ping since I installed them (months ago).
The first RB250GS switch [10.20.2.8] has not missed a single netwatch since I installed it (2 weeks ago).
The devices attached to the second switch have never missed a netwatch ping (2 weeks installed).
The second RB250GS switch [10.20.2.9] loses netwatch many times per day. It sometimes stays down for 2 minutes before recovering.
Try adding static ARP entries on RB450G for both RB250G.
EDIT: 100+ hours and no failures.
This seems to be working. 12 hours and no netwatch failures.
Is this a bug in SwOS?