[ https://issues.apache.org/jira/browse/HDFS-4888?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13693124#comment-13693124 ]
Kihwal Lee commented on HDFS-4888: ---------------------------------- Ravi, here are my review comments: * {{//Automatic safemode}} : the resource-low case is also automatic. Perhaps you meant start-up safe mode? * Removal of the {{datanodeThreshold > 0}} condition: by setting this threshold to 0, the check is disabled. So we still want this check in the original place and also added around the message, "needs additional x live datanodes". * When printing out the extension period, {{reached + extension - now()}} can become negative if repl queue init takes more than the extension (30 seconds by default). Use of Linux THP makes this a lot faster, but it can still exceed 30 seconds for large name spaces. If that happens, the time-to-exit increases every time the message is printed. Rather than Math.abs(), it should say something like "soon". > Refactor and fix FSNamesystem.getTurnOffTip to sanity > ----------------------------------------------------- > > Key: HDFS-4888 > URL: https://issues.apache.org/jira/browse/HDFS-4888 > Project: Hadoop HDFS > Issue Type: Bug > Affects Versions: 3.0.0, 2.0.4-alpha, 0.23.9 > Reporter: Ravi Prakash > Assignee: Ravi Prakash > Attachments: HDFS-4888.patch, HDFS-4888.patch > > > e.g. When resources are low, the command to leave safe mode is not printed. > This method is unnecessarily complex -- This message is automatically generated by JIRA. If you think it was sent incorrectly, please contact your JIRA administrators For more information on JIRA, see: http://www.atlassian.com/software/jira