Management commands available to retrieve internal state from ndb nodes (to start with DUMP command) Remove #ifdef VM_TRACE on certain DUMP commands to make it possible to understand state of a customer system. Direct the trace output to management API so that the trace output can be viewed from command line management server or management client. No fancy stuff, simply output the data as formatted by the kernel. This work log entry is to improve our possibility to find out about internal state of MySQL Clusters with problems. Documentation of the DUMP variants useable is added by those that add DUMP variants useable in a release version. Open up a section in the MySQL Cluster Management Manual for this.