Add a "no-block-delete" mode for use in recovery operations -----------------------------------------------------------
Key: HDFS-2871 URL: https://issues.apache.org/jira/browse/HDFS-2871 Project: Hadoop HDFS Issue Type: Improvement Reporter: Todd Lipcon Occasionally due to operator error or some other reason, the NN metadata is accidentally lost or corrupted. When recovering the cluster, we usually recommend that the operator force the NN into safe mode at startup. But often this isn't quite enough, since they need to exit safe mode in order to start other dependent services like HBase. It would be nice to add a safety config for cases like this, that allows the NN to exit safe mode, but does not issue any block deletions. Thus, if there is some problem, the NN could safely be restarted again with different metadata with no (or minimized) dataloss. -- This message is automatically generated by JIRA. If you think it was sent incorrectly, please contact your JIRA administrators: https://issues.apache.org/jira/secure/ContactAdministrators!default.jspa For more information on JIRA, see: http://www.atlassian.com/software/jira