HDP Cluster Issue : hiveserver2, Datanode and RegiosnServer services keep failing.

[#1]

Today suddenly I noticed I am not able to connect to hive view, on further investigation found that my hiveserver2. Datanode and RegionServer services are keep failing. Here is the detail datanode exception datanodelog.txt

Exception: 2017-01-05 16:54:29,697 INFO common.Storage (Storage.java:tryLock(774)) - Lock on /tmp/hadoop-hdfs/dfs/data/in_use.lock acquired by nodename [email protected] 2017-01-05 16:54:29,701 WARN common.Storage (DataStorage.java:loadDataStorage(449)) - Failed to add storage directory [DISK]file:/tmp/hadoop-hdfs/dfs/data/ org.apache.hadoop.hdfs.server.common.InconsistentFSStateException: Directory /tmp/hadoop-hdfs/dfs/data is in an inconsistent state: Can't format the storage directory because the current/ directory is not empty. at org.apache.hadoop.hdfs.server.common.Storage$StorageDirectory.checkEmptyCurrent(Storage.java:480)

[#2 · Accepted solution]

It would appear that your DataNode is failing which is the cause of the other services failing.

It also appears that you have not changed the default hdfs-site.xml configuration that controls where DataNodes store their data on the local filesystem. It is not uncommon for operation systems to wipe the /tmp directory (on boot). Perhaps you have experienced this and need to re-format your HDFS?

Change dfs.datanode.data.dir, dfs.namenode.name.dir, and dfs.namenode.checkpoint.dir, then format HDFS

$ hdfs namenode -format

Beware: Formatting HDFS is a destructive operation. Do not perform this operation unless all of the data in HDFS is stored elsewhere or can be generated.

[#3]

@Josh Elser Do I need to perform this on all the nodes? For now I am ok if I loose the data as its test cluster.