Replies: 5 comments 6 replies
-
All bindings to transient queues will be deleted when the queues itself are after a node restart. |
Beta Was this translation helpful? Give feedback.
-
Ignoring the actual problem for a moment, can you explain why you decided to design it like this?
|
Beta Was this translation helpful? Give feedback.
-
I faced same problem as #4237. Without node restarts, one of my durable queue in topic exchange stops receiving messages, but on management plugin I can see binding exists, also other queue on different node in same exchange works fine with same routing key (in this case routing key different from queue name). Queue still gets messages from default exchange. Problem solved with unbinding and binding again or with recreating queue. Any ideas what can cause this problem? |
Beta Was this translation helpful? Give feedback.
-
We are facing the same problems as mentioned in #5076 (comment), has anyone found a fix/workaround yet? RabbitMQ version: 3.9.16 |
Beta Was this translation helpful? Give feedback.
-
Adding my own experience, I have seen this happen only twice in the last two years. The latest one was was with 3.12.4 on a 3 node cluster running a combination of quorum and classic mirrored queues. 2 nodes were rebooted which meant the leadership for the queues was concentrated on a single node. Publishers connected to this node experienced 312 NO_ROUTE for two bindings only. When the publisher was connected to another node, it just worked. |
Beta Was this translation helpful? Give feedback.
-
Hi, I have 3 node RabbitMQ cluster (version 3.9.15) running on Kubernetes.
I have topic exchange with 6 classic transient queues binded to it (every queue has 1 routing key,which is equal to queue name). When I restart the node on which some of these queues are located, I recreate them with the same names and same routing keys on another node. Sometimes after restart when I try to publish messages to my exchange I got "Message published, but not routed.", but when I try to publish message using default exchange, everything ok and message is published successfully. This problem can be fixed with queue recreation or with rebinding queue to the exchange.
Beta Was this translation helpful? Give feedback.
All reactions