Details
-
Bug
-
Status: Closed
-
Major
-
Resolution: Fixed
-
None
-
None
-
Mac OSX 10.6.2, Java SE 6 (1.6.0_17)
-
Reviewed
Description
When a simple select query with specified projections is processed, the MapRedTask attempts to serialize the execution plan to a scratch directory on the local file system specified by hive.user.scratchdir configuration property. The default value of this property points to /tmp/hive-${user.name}. On some *nix systems, the /tmp directory gets cleaned up on every reboot thus causing the query execution to fail.
Exception stacktrace
hive> select foo from dummy;
Total MapReduce jobs = 1
Launching Job 1 out of 1
java.io.IOException: No such file or directory
at java.io.UnixFileSystem.createFileExclusively(Native Method)
at java.io.File.checkAndCreate(File.java:1704)
at java.io.File.createTempFile(File.java:1792)
at org.apache.hadoop.hive.ql.exec.MapRedTask.execute(MapRedTask.java:89)
...