[pve-devel] corosync, multicast problem because of vmbr multicast_snooping enabled

Alexandre DERUMIER aderumier at odiso.com
Mon Mar 4 09:39:01 CET 2013


>>That would be great. We need reliable multicast, so that is very important. 
Yes, indeed.


I have found this problem, because last month, I thinked that I was having problem with a custom kernel 3.8,
causing whole corosync cluster goes down.

The problem was more simple, If I simply shutdown this node (node8/8), corosync remove correctly the node from cluster.

But after 2-3 min, something bad happen on network layer, and linux bridges on each hosts doesn't send multicast anymore.
Including linux bridges on other hosts of others proxmox clusters (different multicast address) on same vlan !

Disabling multicast snooping on bridge, or using proxmox host management ip on interface and not on bridge, resolve the problem.

So I really think that they are a bug somewhere in linux bridge multicast_snooping feature.



----- Mail original ----- 

De: "Dietmar Maurer" <dietmar at proxmox.com> 
À: "Alexandre DERUMIER" <aderumier at odiso.com> 
Cc: pve-devel at pve.proxmox.com 
Envoyé: Lundi 4 Mars 2013 06:59:31 
Objet: RE: [pve-devel] corosync, multicast problem because of vmbr multicast_snooping enabled 

> >>The question is if that can have bad side effects? 
> 
> I'll investigate this week with my network engineer coworker . 
> multicast_snooping has been added to linux bridge in 2010. 

That would be great. We need reliable multicast, so that is very important. 



More information about the pve-devel mailing list