Hello, We have a 11 nodes Hadoop cluster installed from Hortonworks RPM doc: http://docs.hortonworks.com/HDPDocuments/HDP2/HDP-2.0.9.1/bk_installing_manually_book/content/rpm-chap1.html
The cluster was working fine since it went on Safe Mode during the execution of a job with this message on the NameNode interface: *Safe mode is ON. Resources are low on NN. Please add or free up more resources then turn off safe mode manually. NOTE: If you turn off safe mode before adding resources, the NN will immediately return to safe mode. Use "hdfs dfsadmin -safemode leave" to turn safe mode off.* The error displayed in the job log is: 2014-08-22 08:51:35,446 WARN namenode.NameNodeResourceChecker (NameNodeResourceChecker.java:isResourceAvailable(89)) - Space available on volume 'null' is 100720640, which is below the configured reserved amount 104857600 2014-08-22 08:51:35,446 WARN namenode.FSNamesystem (FSNamesystem.java:run(4042)) - NameNode low on available disk space. Already in safe mode. On each node we have 5 hdd used for Hadoop And we checked the 5 hdd on the namenode are all full (between 95 and 100%) when the HDFS as still 50% of its capacity available : on the other nodes the 5 hdd are at 30/40% So I think this is the cause of the error. On the NameNode we had some Non HDFS data on 1 hdd, so I deleted them to have 50% of this hdd available (the 4 others are still between 95 and 100%) But this didn't resolve the problem I have also followed the advices found here : https://issues.apache.org/jira/browse/HDFS-4425 And added the following property to the hdfs-site.xml of the NameNode (multiplying the default value by 2) <property> <name>dfs.namenode.resource.du.reserved</name> <value>209715200</value> </property> Still impossible to get out of the safe mode and as log as we are in safe mode we can't delete anything in the HDFS. Is anyone having a tip about this issue? Thankfully, Vincent.
