I have deployed CRDB on a single baremetal and wish to run it in production. I may add additional baremetal nodes when the database grows. It is safe to start like this?
It works just fine but you do not get any of the failover/redundancy that is built into CRDB when using multiple nodes.
Thanks for your reply. So, after a while when the database gets bigger and consume most of the resources, can I just attach another node and start cockroachdb with join ip address? Will the data be replicated automatically without any downtime?
yes, that’s how it works.