The kubernetes cluster with three nodes updates the host configuration node by node. How to ensure that messages are not lost.

[sjkcdpc]

mq.yaml

apiVersion: rabbitmq.com/v1beta1 kind: RabbitmqCluster metadata: name: mq spec: replicas: 3 image: rabbitmq:3.8.7-management rabbitmq: additionalConfig: | default_user=admin default_pass=admin additionalPlugins: - rabbitmq_management - rabbitmq_management_agent - rabbitmq_web_dispatch - rabbitmq_federation - rabbitmq_federation_management - rabbitmq_shovel_management - rabbitmq_shovel override: statefulSet: spec: volumeClaimTemplates: - apiVersion: v1 kind: PersistentVolumeClaim metadata: name: persistence namespace: default spec: accessModes: - ReadWriteOnce resources: requests: storage: 100Gi volumeMode: Filesystem

[Zerpet]

I'm not sure I understand the issue or the question raised.

Can you please describe in details what is happening? What issue are you observing with RabbitMQ?

[sjkcdpc]

Rabbitmq is deployed on the kubernetes cluster with three nodes. The configuration is as above.

Now we need to change the three machines from Intel to AMD. In this process, we worry about rabbitmq data loss. How can we ensure 100% data security?

[Zerpet]

The first recommendation is to upgrade RabbitMQ to a more recent patch. RabbitMQ 3.8.7 is not supported by the Operator, see here: https://github.com/rabbitmq/cluster-operator#supported-versions

Make sure you are using Quorum Queues and read through the reliability guide: https://rabbitmq.com/reliability.html