Bridge Forwarding BPDU (STP) Spanning Tree in Cluster Problem

  • Thread starter Thread starter h41cyon
  • Start date Start date
H

h41cyon

Guest
I got two proxmox machines working together in a cluster. They are currently connected to a switch, which has an uplink to router that monitors for redundant ports using the Spanning Tree Algorithm.

Problem is, if I reboot one of my proxmox machines, when it boots up, it sends out a bridge forwarding packet that goes all the way up through the switch and into the router. The router sees this and then promptly shuts down that port for a short period of time.

Question is, why are my two proxmox machines sending these packets out? Is there anyway to stop this from happening? Do you think the issue would be resolved if I moved one of the proxmox machines from behind the switch to a different port on the router?

Any help is appreciated!
 
this is strange, stp is normally disable on bridges.

can you do a "brctl show", it'll display the bridge list with STP values.

they are also some variables tunnables in
/sys/class/net/vmbr0/bridge/....
 
If I run brctl show on both machines, I see that STP is disabled. So that looks fine, but I think it has more to do with the cluster service. I saw something mention using unicast instead of multicast so I tried the following:


  • add the new transport="udpu" in /etc/pve/cluster.conf.new

<cman keyfile="/var/lib/pve-cluster/corosync.authkey" transport="udpu"/>


Did a service pve-cluster restart and WHAM!, port got shut down. Not sure what else to try.
 
I believe STP is what's causing it. I'm not entirely sure, I need to check still. Either way, the issue seems to be with cluster to cluster communication.
 
You must log in or register to reply here.
Top Bottom