No good epoch log available on node

Error Message

Explanation

There are a number of possible reasons for this error message, including an abnormal startup or shutdown. Every node in the cluster must be started with the same recovery epoch. Non-matching recovery epochs occur when a cluster has experienced an unsafe shutdown.

Workaround

1. Make sure that all nodes are powered on.
2. Start the database again.
3. Make sure that all nodes have Spread running. If necessary, restart Spread where it is not running and start the database.
4. On each node that did not start up, examine dbLog for the cause of the failure.
5. If the cause cannot be determined, it's likely that a node has no catalog version or epoch log from which to recover. Run diagnostic tests (see Using Diagnostic Tools) and contact Technical Support.

   To submit problem reports, questions, comments, and suggestions, please use the Technical Support page on the Vertica Systems, Inc. Web site:

   http://www.vertica.com/support

   You must be a registered user in order to access the page.

   Before reporting a problem, please run the Diagnostics Utility described in the Troubleshooting Guide and attach the resulting .zip file.