Page 1 of 1

Strange LACP behavior ?

Posted: Mon Jun 24, 2019 3:33 pm
by efbertos
Anyone have any issues with lacp behavior on nxos w/ vPC ?? i.e...


SW35_lab(config-if-range)# 2019 Jun 24 10:46:04 SW35_lab %$ VDC-1 %$ %LACP-5-LACP_SUSPEND_INDIVIDUAL: LACP port Ethernet1/53(0x1a006800) of port-channel Ethernet1/53(0x1a006800) not receiving any LACP BPDUs suspending (individual)



I was able to get things to work somewhat smoother for a time by using "lacp rate fast" which isn't recommended on vPC peer-links. When I started adding a couple more port-channels i ended up running into this error:


Until I ran into this error while running back-to-back vPC:
SW34_lab %$ VDC-1 %$ last message repeated 2 times
SW34_lab %$ VDC-1 %$ %LACP-3-SYN_COLL_DIS_EN: WARNING: Potential Interop issue on [Ethernet1/53(0x1a006800)]: SYNC, COLLECT and DISTRIBUTE flags enabled too early by partner
SW34_lab %$ VDC-1 %$ last message repeated 2 times
SW34_lab %$ VDC-1 %$ %LACP-3-SYN_COLL_DIS_EN: WARNING: Potential Interop issue on [Ethernet1/54(0x1a006a00)]: SYNC, COLLECT and DISTRIBUTE flags enabled too early by partner
SW34_lab %$ VDC-1 %$ last message repeated 2 times
SW34_lab %$ VDC-1 %$ %LACP-3-SYN_COLL_DIS_EN: WARNING: Potential Interop issue on [Ethernet1/53(0x1a006800)]: SYNC, COLLECT and DISTRIBUTE flags enabled too early by partner


I think the solution is to switch to channel-group x mode on which is non-lacp, but when I am modifying my config to be something it wouldn't be real-world..

Is it possible we're running into a linux kernel issue? i.e. "Bug 202663 - LACP protocol in bridge not forwarding"
https://bugzilla.kernel.org/show_bug.cgi?id=202663


Thanks,

Re: Strange LACP behavior ?

Posted: Mon Jun 24, 2019 8:25 pm
by efbertos
Anyone ?

Re: Strange LACP behavior ?

Posted: Tue Jun 25, 2019 8:56 am
by Uldis (UD)
As the virtual nodes has no L1, try shut no shut interfaces

Re: Strange LACP behavior ?

Posted: Wed Jun 26, 2019 2:11 pm
by efbertos
Uldis (UD) wrote:
Tue Jun 25, 2019 8:56 am
As the virtual nodes has no L1, try shut no shut interfaces
You're right, it was a L1 issue!

I still have one port that goes suspended but at least it's mostly working now thanks