Member since
05-20-2016
14
Posts
2
Kudos Received
1
Solution
My Accepted Solutions
Title | Views | Posted |
---|---|---|
301 | 02-16-2017 10:54 PM |
03-27-2017
07:35 PM
@Dylan Wu Can you paste the error stacktrace from hive.log?
... View more
03-23-2017
03:36 AM
@Geoffrey Shelton Okot You can follow instructions similar to this: https://docs.hortonworks.com/HDPDocuments/HDP2/HDP-2.4.2/bk_Security_Guide/content/_set_up_truststore_for_ambari_server.html. Basically you'll need to import certificates to your truststore/keystore.
... View more
03-23-2017
03:29 AM
@Preeti Padelkar Can you post the error trace from client (& server if you have one)?
... View more
03-23-2017
03:24 AM
@cmcbugg You can use the following jvm options to add additional logging for Kerberos: -Dsun.security.krb5.debug=true -Dsun.security.jgss.debug=true -Djavax.net.debug=all When you run your kafka process at DEBUG log level, it should ideally log the kerberos ticket exchange, which will contain the KDC coordinates. Hope this helps
... View more
03-23-2017
03:21 AM
@Cruz DSouza It appears that HBase is not in the classpath and that's why the class not found error. Can you try adding hbase jars to your classpath and try again?
... View more
03-23-2017
03:17 AM
@Marcy For earlier hive versions (and if you are not running Ranger), you can try SQL Standard Auth: https://cwiki.apache.org/confluence/display/Hive/SQL+Standard+Based+Hive+Authorization. However, it is preferred to use Ranger for fine grained authorization. You can find details here: https://docs.hortonworks.com/HDPDocuments/HDP2/HDP-2.3.2/bk_Sys_Admin_Guides/content/ref-746ce51a-9bdc-4fef-85a6-69564089a8a6.1.html https://docs.hortonworks.com/HDPDocuments/HDP2/HDP-2.3.2/bk_installing_manually_book/content/installing_ranger_plugins.html Hope that helps
... View more
03-23-2017
03:09 AM
@Morten R. What do you see in HS2 logs and also the application logs? Can you change the log level to DEBUG and then try to get the error stacktrace?
... View more
03-23-2017
03:03 AM
1 Kudo
@Ye Jun Hive web interface (HWI) is a legacy component, which will be removed in next releases from Apache. If you don't plan to use Ambari views, you can use the HiveServer2 Web UI as described here: https://cwiki.apache.org/confluence/display/Hive/Setting+Up+HiveServer2#SettingUpHiveServer2-WebUIforHiveServer2. Hope that helps.
... View more
02-16-2017
10:54 PM
1 Kudo
Hi @Davide Vergari, If you ask beeline to run commands in one file, they will run sequentially. However, you could try the following: 1. Split the queries in different files and run multiple beeline commands in different threads in your program. 2. Use JDBC where you can open multiple connections in different threads and run queries on each connection. Hope that helps.
... View more
02-16-2017
10:49 PM
@jzhang The _HOST in the principal name gets replaced by the fully qualified domain name of the host at runtime. However, this needs the reverse DNS to be configured correctly on the box. Could you check that?
... View more
02-16-2017
10:41 PM
It would be good to look at the app logs for the query to see where the tasks are stuck. Some helpful tips to debug tez: https://cwiki.apache.org/confluence/display/TEZ/How+to+Diagnose+Tez+App
... View more
12-09-2016
09:38 PM
@Andrew Sears Can you get the output of jstack when this happens? The threads go back to the threadpool once a query has completed execution, so unless there is some kind of a leak, this should not happen. The number of queued tasks also seem pretty huge.
... View more
12-09-2016
09:27 PM
HiveServer2 and Metastore servers can be on different machines. HS2 can also be configured to get metadata via a remote Metastore server (HS2 opens a TCP client connection to Metastore server) or use an embedded Metastore server (in which case HS2 calls the Metastore Server's API directly: use hive.metastore.uris="" in hiveserver2-site.xml to achieve that).
... View more
05-20-2016
04:21 PM
@Christian Guegi Do you see this issue only when running multiple worflows in parallel?
... View more