No Multicast Address? - Can't get key totem.interface.0.mcastaddr. Error CS_ERR_NOT_EXIST

qubozik

Member
Jun 4, 2016
16
0
21
39
Hello!

I have a two node cluster.

NodeA in the cluster does not appear to have a multicast address. When running the following command this is the output:

# corosync-cmapctl -g totem.interface.0.mcastaddr
Can't get key totem.interface.0.mcastaddr. Error CS_ERR_NOT_EXIST

NodeB will show a multicast address when running this command.

NodeB was joined into NodeA when the cluster was created..

Also, when using the web GUI, you can manage the entire cluster from NodeA without issues, but when using the web GUI when logged into NodeB, you cannot manage NodeA. You will see various loading screens that eventually time out.

It does not appear that multicast is working. Running test with omping and the like show errors. I believe this may be why.

Any way to force NodeA to pull a multicast address? Or, am I understanding something incorrectly?

This error was mentioned before in a thread. The OP eventually upgraded his cluster to the next .X version and the error went away. No such luck in my case.

One item to mention.. I have never rebooted NodeA after the cluster was created. There are production VMs running on this machine, so reboots are something to be prevented at all costs whenever possible.

Any ideas? Thank you in advance.

BTW.. I have looked into Multi-cast related items and have the following information:

- My switch supports Multicast and IGMP snooping, but does not support the querier service.
- I enabled the querier on the Linux Bridge as outlined in the Multicast Notes Proxmox KB. Not sure if this is the reason? I thought NodeA should still have an address.

If multicast is the problem, I could switch to unicast for the cluster, but I would like to find a way to do it that would not require a reboot if possible.
 
Just wanted to follow-up on this in case anyone has the same issues...


So, the multicasting issues were caused by the enabled "Anti-DDoS" features which were enabled within the switch. I disabled those and was then able to use omping to test multi-cast between the nodes.

Something still wasn't right as nodeA was not reporting having a multi-cast address just as it was previously.

As I mentioned, downtime would be an issue, so I left it the way it was.

An unrelated issue occurred which forced a reboot of nodeA. Upon coming back online, all cluster functionality was working as it should.

So, I believe the issue was caused initially by creating the cluster while having the anti-ddos features enabled on the switch. This was causing issues with cluster communication. Not sure exactly which setting caused this as there are a handful. I just disabled all of them and now everything works as it should.

Perhaps I will try and correlate which setting at a later time.

Hopefully this helps someone as it was a configuration issue that doesn't jump right out at you at first.
 

About

The Proxmox community has been around for many years and offers help and support for Proxmox VE, Proxmox Backup Server, and Proxmox Mail Gateway.
We think our community is one of the best thanks to people like you!

Get your subscription!

The Proxmox team works very hard to make sure you are running the best software and getting stable updates and security enhancements, as well as quick enterprise support. Tens of thousands of happy customers have a Proxmox subscription. Get yours easily in our online shop.

Buy now!