[PVE-User] Whole cluster brokes

[Thomas Lamprecht]

Hi,
On 03/08/2017 11:38 AM, Daniel wrote:
> Hi,
>
> when i try the command with 2 NODES i got the follwing Error.
> So it seems realy to be a multicast problem.
>
> root at host01:~# omping -c 10 -i 1 -q 10.0.2.110 10.0.2.111
> 10.0.2.111 : waiting for response msg
> 10.0.2.111 : waiting for response msg

Command is ok like this, thje one from your other mail is not.
But you have to start it on both the node with IP 10.0.2.110 *and* the 
one with 10.0.2.111 to make it work.

>
> I cant restart pve-cluster – I get errors. Corosync was not restarted yet – And yes – actually I don’t have HA configured yet.
> Is there any special command to restart Corosync?

systemctl restart corosync

>
> Should this help when I try to do on one node?
>
> echo 1 > /sys/devices/virtual/net/vmbr0/bridge/multicast_querier
>   

Yes, you can try that.

> I am not sure what how long the cluster was working after 13 was shutdown.
>
Changes on the switch/network?

> ok it seems that Multicast is not working anymore. But how can this happen? It was working before without any trouble.

As said, or changes in the network or that the other node really acted as a
multicast querier.

But omping looks like no multicast is working at all, with a missing 
querier you
would get problems after about 5 minutes but before that it should work.