The orphaned sqoop task issue described in SQOOP-1916 <https://issues.apache.org/jira/browse/SQOOP-1916> was observed in production on sqoop 1.99.5 and 1.99.6 (no other releases were tested). This issue became a problem because each orphaned sqoop process would consume a fraction of host memory, resulting in host memory saturation over time. The JVM stack dump posted to the ticket is identical to the stack dump of an orphaned task on the yarn cluster.
A patch containing a fix has been attached to SQOOP-1916. This is a heads-up in case anyone else has encountered this issue. I'm not sure who's alerted by the JIRA issue, so I'm casting a wider net by posting to this group.
