If we kill “enough” nodes of a cluster (let’s say 3-node cluster and 2 nodes are down), any interaction attempt with the remaining node just hangs. This includes SQL transactions and administrative access (CLI or Web UI). Is this by intentional design or something you plan to change in the future?
** Note: this is a copy of a community asked question on a different communication channel
Yes, this is a known problem with our current design, and we do want to fix it. However, there is only so much we can do to fix this - if a node is crashing, there’s not much CockroachDB’s built-in monitoring can do to tell you what’s going on with that node. When too many nodes are failing, you’ll generally need to rely on outside-of-CockroachDB tools to bring them back up. We have an open issue to make the transaction timeout period for SQL transactions configurable, but you can’t do this yet.