Details
-
Improvement
-
Status: Open
-
Major
-
Resolution: Unresolved
-
None
-
None
-
None
-
None
Description
When running in dynamic environments like EMR where the cluster that Drill is running on is transient the ip address of the Hadoop name node is different from one day to the next. This makes it very difficult to configure a storage plugin/workspace that points to the HDFS file system. Everytime we startup a new EMR cluster we have to manually open the dfs storage plugin config and change the ip address in the hdfs:/// connection field to the currently running instance. It would be very useful to be able to have Drill figure out this IP address itself.