Occasionally, things go terribly wrong and the extent map gets out of sync with the data. This might happen, for instance, when a node goes offline during an import.

We need a utility to re-create the extent map based on the data. The following are ideas that may apply.

This is an emergency process, not to be used in normal day to day activities.

It will be slow. While rebuilding, all processing for any table being worked on will be blocked, including reads. All current queries must be completed before starting work.

There should be an option to select the tables/databases to be worked on.

Do we want sequential (one table at a time) or parallel processing?