kubernetes-peer discovery using hostname/nodename

[kk-manh]

Hi, Can you provide an example using the hostname rather than IP address. Using an IP address in a fluid environment like K8S where pods can be killed / spun up any number of times with different IP addresses, using IPs are inherently bad. Especially when you need backup/restore from persistnt volumes and you end up with a bunch if folders in the mnesia folder that all have old / stale IP addresses.

[lukebakken]

Thank you for your time and for using RabbitMQ.

The RabbitMQ team uses GitHub issues for specific actionable items engineers can work on. This means that GitHub issues are not used for questions, investigations, root cause analysis, discussions of potential issues, and so on.

I see the following rabbitmq-users message. Please post this question as a new discussion on the mailing list.

Thanks!

[kk-manh]

This is an issue. because the rabbitmq server and the peer discovery mechanism fails when the HOSTNAME is used for the clustering mechanism.

[lukebakken]

@kk-manh - you're asking for an example, which isn't considered to be an issue by the RabbitMQ team. We're happy to follow up on the mailing list. If we determine this to be an issue or new feature worth further investigation, a GitHub issue will be created at that time.

Did you read the documentation? Specifically, this section that mentions the use of hostnames:

When a list of peer nodes is computed from a list of pod containers returned by Kubernetes, either hostnames or IP addresses can be used. This is configurable using the cluster_formation.k8s.address_type key:

[daudn]

Did anyone find a solution for this?

[lukebakken]

@daudn - Please either see the examples or ask a question on the mailing list.