@RunLLM If I deployed a Pulsar cluster in Kubernetes with only one zookeeper, is there any safe way to reconfigure the cluster to have multiple zookeepers?

It seems likely that you can safely scale from a single ZooKeeper node to a multi-node ensemble, but you must follow ZooKeeper's quorum reconfiguration process. You need to update the

zookeeper.conf

to add new server entries, assign unique IDs, and ensure each node has the correct

myid

file. After updating configs and starting new ZooKeeper pods, the ensemble will form a quorum. This process is described for bare metal, but the same principles apply in Kubernetes, though you must ensure persistent storage for each ZooKeeper pod and update StatefulSet or Helm values accordingly. There is no explicit Pulsar-specific documentation for this migration, so care is needed to avoid data loss or split-brain scenarios. See the official ZooKeeper multi-server setup guide for details. Would you like more detail on the step-by-step process for Kubernetes deployments? Sources: - Deploy local ZooKeeper 👍 0 button 👎 0 button 💡 Hint: Mention @runllm in the thread for followups.

@RunLLM do i need to update a configuration to make the zookeeper create a quorum when it was originally in standalone mode, or will it automatically pick up the new zookeepers and update itself?

You must update the ZooKeeper configuration (

zookeeper.conf

) on all ZooKeeper nodes to include the new servers; ZooKeeper will not automatically detect or form a quorum with new nodes unless explicitly configured. Each node needs the correct

server.N

lines and a unique

myid

file. After updating configs and starting the new nodes, the ensemble will form a quorum as described in the multi-server setup guide. Would you like a step-by-step example of the configuration changes? Sources: - Deploy local ZooKeeper 👍 0 button 👎 0 button 💡 Hint: Mention @runllm in the thread for followups.