We have some CRS deployed in our network. One of them works unstable. After running for some days (sometimes a week, sometimes a few hours) it suddenly drops throughput rate to about 50%, and packet loss at about 10 - 40% depending on packet size.
Simple diagram: (all in layer 2, 802.1q tags, traffic volume = about 700 Mbit/s)
This packet loss can be observed between A and B, but pinging CRS from A or from B works ok. There are no errors in log, cpu is at 3%, only reboot helps.
Code: Select all
[ A ] === [ CRS ] === [ B ]
ROS v6.35.4 (already updated), firmware 3.23
Any ideas where to look for solution? Seems to be something wrong on the switch side, thus I have no idea how to diagnose it.