Issues Fixed in CDH 5.1.0
The following topics describe known issues fixed in CDH 5.1.0.
The same DataNodes may appear in the NameNode web UI in both the live and dead node lists
YARN Fair Scheduler's Cluster Utilization Threshold check is broken
Workaround: Set the yarn.scheduler.fair.preemption.cluster-utilization-threshold property in yarn-site.xml to -1.
ResourceManager High Availability with manual failover does not work on secure clusters
Workaround: Enable automatic failover; this requires ZooKeeper.
MapReduce over HBase Snapshot bypasses HBase-level security
The MapReduce over HBase Snapshot bypasses HBase-level security completely since the files are read from the HDFS directly. The user who is running the scan/job has to have read permissions to the data and snapshot files.
Workaround: MapReduce users must be trusted to process/view all data in HBase.
HBase snapshots now saved to the /<hbase>/.hbase-snapshot directory
HBase snapshots are now saved to the /<hbase>/.hbase-snapshot directory instead of the /.snapshot directory. This was a conflict introduced by the HDFS snapshot feature in Hadoop 2.2/CDH 5 HDFS.
Workaround: This should be handled in the upgrade process.
Oozie jobs don't support ResourceManager HA in YARN
If the ResourceManager fails, the workflow will fail.
Oozie HA does not work properly with HCatalog integration or SLA notifications
This issue appears when you are using HCatalog as a data dependency in a coordinator; using HCatalog from an action (for example, Pig) works correctly.
|<< Issues Fixed in CDH 5.1.2||Issues Fixed in CDH 5.0.6 >>|