I have two independent rabbitmq clusters (both with v3.6.15), broken down from a older and bigger cluster. AFAIK, the config from the old cluster was exported and re-imported on both clusters and created several shovels to import the old cluster data. Clusters where manually created, no auto-discovery exist.
Everything seems to work from at least one year... but when i enter cluster1 and check the queues, it shows queues from cluster2. Doing the same on cluster2 i only see data from that cluster, no info from cluster1.While i have some federations and some shovels to other third party rabbitmq, none are cross cluster.
After some debug, i found that cluster1 still have some vhosts that where migrated to cluster2 and looks like cluster1 is mixing info from it own queues with info from cluster2. Why the reverse also do not happen is unknown for me.
I created a iptables on cluster1 to block traffic between the 2 clusters and it "solved" the problem. I then found in the logs this message with reference to the cluster2:
Management delegate query returned errors
So the question is, what is this management delegate and how to control it?
How to fix this problem without using a firewall.
0 Answers