Issue between MaxScale galeramon and Galera arbitrator

[Tobia Scapin]

Hello, I have a issue with my configuration and maxscale.

I have a galera cluster of 2 nodes and 1 galera arbitrator.

Maxscale 24.02 is installed on both nodes with galermon and rwrouter with root_node_as_master option.

Today I had to stop one of two nodes, the master one, and maxscale didn't failover to the second node. It's status was "slave" and it was not promoted. 

After a lot of bad words I discovered that the unique alive node had  wsrep_local_index=1 and probably due to root_node_as_master settings this was the reason of the missing failover.

Then I started to seach why its index was 1 and I discover that it was assigned to the galera arbitrator.

Stopping the arbitrator the failover was made as expected.

Do you think this is an issue or I have a wrong configuration? It is normal in your opinion that an arbitrator can take the 0 index?

Thanks for your opinion.