Hey People!
So I have these switches in a stack and I have just patched in a physical Centos 7 server using 3 uplinks. 2 uplinks go into one switch and 1 uplink goes into the second switch in the stack. Basic access ports allowing that vlan. Switches are running an older IOS 15.0(2) SE4
I have configured teaming on the nic using roundrobin (which is new to Centos7), when the server is in use under load, we seem to get a lot of mac address flapping between ports.
The Linux teams virtual mac is flapping between all ports and on top of that memory gets low on the switch and it eventually just reboots itself.
Now I am happy to configure Linux to use activebackup configuration, but I liked the idea of all 3 links being active, but I don't understand why it is causing the switch to reboot?
We disabled one port and then it just flapped between the 2 remaining uplinks and finally when we disabled the 2nd port on the switch the flapping stopped...well because it had no where else to go heh
I am not aware that the roundrobin config in centos 7 requires any kind of special physical switch config?
%SW_MATM-4-MACFLAP_NOTIF: Host VIRUTAL TEAM MAC ADDRESS in vlan XX is flapping between port Gi2/0/42 and port Gi1/0/41
If anyone has any input or info or can even point me towards a possible cause or anything, I'd be grateful!
Cheers
No comments:
Post a Comment