I am using storm 0.9.3 and trying to make use of the HiveBolt to stream the data directly into hive tables.
I am following this example: https://github.com/hkropp/storm-hive-streaming-example But instead of a kafkaspout I am using a random data generating spout to keep things simple. When I submit the topology, the bolt thread is able to connect to my hive-metastore but then fails with these errors once it was: *" Non-local session path expected to be non-null"* and then :* "Failed connecting to EndPoint hive " * I also made sure that the table is already created and accessible. I also tried it on partitioned and un-partitioned tables. Any help as to what is going wrong? Regards, Sunit On Wed, Mar 18, 2015 at 9:51 AM, Sunit Swain <[email protected]> wrote: > I am using storm 0.9.3 and trying to make use of the HiveBolt to stream > the data directly into hive tables. > > I am following this example: > https://github.com/hkropp/storm-hive-streaming-example > > But instead of a kafkaspout I am using a random data generating spout to > keep things simple. > > When I submit the topology, the bolt thread is able to connect to my > hive-metastore but then fails with these errors > once it was: *" Non-local session path expected to be non-null"* > and then :* "Failed connecting to EndPoint hive " * > > I also made sure that the table is already created and accessible. I also > tried it on partitioned and un-partitioned tables. > > Any help as to what is going wrong? > > Regards, > Sunit >
