Startup Phases

While your storage nodes are starting up, you can get slightly more detailed information about what they are doing via the management client.

If you enter STATUS, the management client will tell you what phase of the startup process you are in, as in this example:

ndb_mgm> 2 STATUS Node 2: starting (Phase 2) (Version 5.0.11) ndb_mgm> 2 STATUS Node 2: started (Version 5.0.11)

You can also use the ALL STATUS command to list the status for all nodes at the same time.

A quick description of what the cluster is doing during each phase follows; this is often useful when you are debugging problems as it gives you a hint as to exactly what problems you are experiencing:

Note

Note that some concepts here have not been explained in detail yet; they are intended to help you troubleshoot your problems, but if you do not have any problems yet or do not follow any of the points, feel free to skip this.

• Phase 1 is when the node first connects to all the other nodes in the cluster and heartbeats between nodes are initiated.
• Phase 2 is when an arbitrator is elected, if none exists.
• Phase 3 initializes the internal variables.
• Phase 4 finds the end of the logs and prepares for the node to start.
• Phase 5 copies the data from the alive nodes in the node group (if available) or starts the recovery from local checkpoints and the REDO log.
• Phase 6 and onward do various small other tidbits such as re-creating all ordered indexes, so you can expect high CPU usage.