Loss of a single LACP link results in LACP group traffic failing


hi


can explain one?

have following setup:

- switches:

    3 x cisco 3750x-24ts , 1 x 3750g-24ts-1u in stack
    s/w: 15.0(2)se on all
    etherchannel groups created lacp/active 4 6 links per group


- servers #1

    4 x dell servers broadcom nics in hyper-v 2012 r2 cluster
    nic teaming configured using ms hyper-v load balancing (i.e. not using broadcom utility aggregate nics).
    there 6 nics per lacp group on each server connected 3750(x) switch stack.
    nics connected across 4 members of stack. lag created lacp/active

- servers #2

    2 x dell servers intel nics running windows 2012 r2 in cluster (used ha file cluster).
    nic teaming created using win2012 load balancing. lag created lacp/active
    there 4 nics per lacp group on each server connected 3750(x) switch stack
    nics connected across 4 members of stack. lag created lacp/active


issue:

on thursday helpful telecoms engineer knocked out power switch #2 of 3750(x) core switch stack , caused reboot of switch. when happened appears that, rather losing connectivity on lacp links connected single switch, entire lacp interface went down causing both hyper-v cluster , ha file cluster start throwing roles around , denying existence of other nodes. chaos ensued.

once switch #2 had rebooted joined stack correctly , normal service, in terms of network connectivity resumed. file cluster started operating without assistance of vms had rebooted , in various states of chaos.

question have happened lacp lag interface on switches or loadbalancing on hyperv cause entire group fail when 1 / 2 links went down?

in advance - greatfully received!
 
martin

uepals ms forum data

leo

thanks reply.  posted on cisco forum (i'm running 3750x stack) @ same time , today there has pointed me possible issue if it's stack master gets removed stack whatever reason lacp connections can dropped mac address of stack changes. have recommended running following config command on stack:  stack-mac persistent timer 0

i don't think it'll broadcom nics (although have had fun them in win2012r2 hyperv servers) i've had problem intel nics. if switch config change doesn't work i'll give dell shout.

thanks again

martin


uepals ms forum data



Windows Server  >  Hyper-V



Comments

Popular posts from this blog

Motherboard replacement

Cannot create Full Text Search catalog after upgrading to V12 - Database is not fully started up or it is not in an ONLINE state

Remote Desktop App - Error 0x207 or 0x607