Glad that you tried resetting the property to check the actual issue. I think the actual issue was, the way you tried to access the hive entities from UI. You could have tried DSL search like below. ... View more

@Floris Smit From the atlas application log, the metadata is consumed by atlas and everything looks fine.. 2016-10-12 10:05:16,487 INFO - [NotificationHookConsumer thread-0:] ~ Updated entities: { "entities": { "updated": [ "52bf585a-8e66-4402-864a-84bc201daffc", "f67d81d9-4e2e-4396-99cf-2f2166e9fd83", "43c64fde-38c6-4dbc-b75e-86d804e619c0", "f5cdd132-0066-4a37-82a3-837d9316650d", "53015e6b-2588-4c3c-925b-1cc2596eebc5", "3ae5d014-147e-468d-a442-e4253d647fdc", "ae0dacc8-66ac-4c5f-8673-35a3d1ca678c", "fdc38fcc-edb9-48f6-a2f8-282a9032793a", "c33fc14c-7aeb-45cb-b933-8e3c3d9f5a6e", "52c78367-162c-4e91-b5f1-5335c8f2ad16", "59af504a-192c-4814-b657-80e3c3a5bb8b", "367ea90f-4bb7-4fc8-a43a-0fba1250cd29" ] } } (EntityResource:210) Even the hive logs reflect the same. Also, atlas.hook.hive.synchronous is setting used to publish messages at the user convenience(sync/async). In your case, even after setting atlas.hook.hive.synchronous to false, the metadata should have been published to atlas. atlas.hook.hive.synchronous - boolean, true to run the hook synchronously. default false. Recommended to be set to false to avoid delays in hive query completion. source: http://atlas.incubator.apache.org/Bridge-Hive.html ... View more

@Floris Smit Can you please share the snapshot(attach the log) of the atlas application log? Also, try creating a sample table in hive and capture the hiveserver2.log and atlas application log for the sample table creation time. Please paste this log as well. Above logs are needed for debugging. Also, please check if all the dependent services are up and running(HBase(regionservers), Kafka, Ambari-infra). ... View more

@Mohan V Not sure what went wrong while creating the topic. It could be related to disabling kerberos, as this might have some stale bits. Can you check if the topic is actually created using kafka command line tools? Also, try creating the topic using the FQDN(zookeeper quorum) for zookeeper instead of localhost. for example: <strong>./kafka-topics.sh --create --zookeeper <host1:port>,<host2:port> --replication-factor 1 --partitions 1 --topic test</strong> If none of the above helps, can you try deploying a new cluster without kerberos if possible? ... View more

@Mohan V Can you please provide more details on how you created the topic(the exact command)? Do you have permissions to access that topic? Is ranger enabled on the cluster? - if yes, please check if you have policy for enabling permissions to create kafka topic. Is the cluster kerberized? If yes - Are you performing the producer operations as a proper user with proper 'kinit'? There is a similar question asked in the community, please check this also if it helps(https://community.hortonworks.com/questions/42793/kafka-producer-error-no-partition-metadata-for-top.html).. ... View more

@Michael Young Thanks for sharing your findings, this is helpful. ... View more

@Sundar Lakshmanan Is this issue resolved? ... View more