I have a fabricpath network which consists of Nexus 5500 switches and fexes. Most of my FEXes are dual homed between two 5500 switches. This is working well except for one issue that has come up occasionally.
One of my FEXes has some ESX servers connected to it. Occasionally I get certain error messages like:
FWM-2-STM_LOOP_DETECT: Loops detected in the network for mac xxxx.xxxx.xxxx among ports gpc1.aaa and gpc1.bbbb
aaa is the emulated switch-id and bbbb is the switch-id for one of the switches in the VPC. While looking into it, I was seeing that a lot of mac addresses seemed to be in the same boat (show mac address on a different switch was showing either aaa or bbb for the SWID).
I was under the impression that the VPC would be sending the shared switch-id for all mac addresses.
I'm having a lot of trouble figuring out how to even search for this issue. Any suggestions as to what might be causing it?
My problem turned out to be a VPC+ issue that occurred after a link to a FEX module flapped
%FWM-2-STM_LOOP_DETECT: Loops detected in the network for mac 0015.c5eb.1742 among ports gpc1:302 and gpc1:1301 vlan 148 - Disabling dynamic learn notifications for 180 seconds
The loop message referenced the ID of the virtual switch (1301) and one of the physical switches (302). A shut/"no shut" on the port channel on the physical switch got it back to the right state.
is this issue being resolved?
i have a similar issue in my fabricpath network
except that I don’t use vpc+.
and I recently see this loop detect log frequent that impacting the network performance.