Member since
09-29-2015
122
Posts
159
Kudos Received
26
Solutions
My Accepted Solutions
Title | Views | Posted |
---|---|---|
6716 | 11-12-2016 12:32 AM | |
1925 | 10-05-2016 08:08 PM | |
2640 | 08-02-2016 11:29 PM | |
23357 | 06-24-2016 11:46 PM | |
2061 | 05-25-2016 11:12 PM |
11-22-2017
06:46 PM
Can you turn up Livy log level and post the log in the livy-server?
... View more
11-17-2017
06:16 PM
Can you share the details of how Livy job is launched? This exception appears to come from Livy side, did the Spark job actually get launched? What are the errors on Spark executor side?
... View more
10-13-2017
07:14 PM
7 Kudos
Zeppelin Best Practices
Install & Versions
Leverage Ambari to install Zeppelin and always use the latest version of Zeppelin. With HDP 2.6.2, Zeppelin 0.7.2 is available and it contains many useful stability & security fixes that will improve your experience. Zeppelin in HDP 2.5.x has many known issues that were resolved in 2.6.2 Deployment Choices
While you can select any node type to install Zeppelin, the best place is a gateway node. The reason gateway node makes most sense is when the cluster is firewalled off and protected from outside, users can still see the gateway node. Hardware Requirement
More memory & more Cores are better Memory: Minimum of 64 GB node Cores: Minimum of 8 cores # of users: A given Zeppelin node can support 8-10 users. If you want more users, you can set up multiple Zeppelin instances. More details in MT section. Security: Like any software, the security depends on threat matrix and deployment choices. This section assumes a MT Zeppelin deployment.
Authentication
Kerberize HDP Cluster using Ambari Configure Zeppelin to leverage corporate LDAP for authentication Don’t use Zeppelin’s local user based authentication, except for demo setup. Authorization
Limit end-users access to configure interpreter. Interpreter configuration is shared and only admins should have the access to configure interpreter. Leverage Zeppelin’s shiro configuration to achieve this. With Livy interpreter Spark jobs are sent under end-user identity to HDP cluster. All Ranger based policy controls apply. With JDBC interpreter Hive & Spark access is done under end-user identity. All Ranger based policy controls apply. Passwords:
Leverage Zeppelin’s support for hiding password in Hadoop credential for LDAP and JDBC password. Don’t put password in clear in shiro.ini Multi - Tenancy & HA
In a MT environment, only allow admin role access to interpreter configuration A given Zeppelin instance should support only < 10 users. To support more users, setup multiple Zeppelin instance and put a HTTP proxy like NGinx with sticky sessions to route same user to same Zeppelin instance. Sticky sessions are needed since Zeppelin stored notebook under a given Zeppelin instance dir. If you use a networks storage system, the Zeppelin notebook directory can be stored on the network storage and in that case sticky sessions are not needed. With upcoming HDP 2.6.3, Zeppelin will store notebooks in HDFS and this requirement will not be necessary. Interpreters
Leverage Livy interpreter for Spark jobs against HDP cluster. Don’t use Spark interpreter since it does not provide ideal identity propagation. Avoid using Shell interpreter, since the security isolation isn’t ideal. Don’t use the interpreter UI for impersonation. It works for Livy & JDBC (Hive) and for all others we don’t officially support i Users should restart their own interpreter session from the notebook page button instead of the interpreter page which would restart sessions for all users Livy interpreter JDBC interpreter Also See Jianfeng Zhang's Zeppelin Best Practices notebook
... View more
Labels:
09-12-2017
10:00 PM
This is coming very late but it appears you are launching a Spark Standalone cluster and not really submitting Spark on YARN. For Spark on YARN, as per http://spark.apache.org/docs/latest/running-on-yarn.html the spark-submit needs to say --master yarn. The actual details of yarn cluster details are picked up from yarn conf dir. ./bin/spark-submit --class path.to.your.Class --master yarn --deploy-mode cluster [options] <app jar> [app options]
... View more
06-29-2017
06:24 PM
You may want to look at Spark HBase connector that is based on DataFrame and should give better performance. https://github.com/hortonworks-spark/shc
... View more
05-25-2017
09:46 PM
This is applicable in a Kerberos enabled HDP 2.5.x cluster with Zeppelin, Livy & Spark. Post successful Kerberos setup, log in to Zeppelin and run Spark note, the note runs file. But running simple sc.version from livy interpreter gives "Cannot start spark" in the Zeppelin UI. In the Livy log at /var/log/livy/livy-livy-server.out you may find a message similar to the following. INFO: 17/05/25 21:24:12 INFO metastore: Trying to connect to metastore with URI thrift://vinay-hdp25-2.field.hortonworks.com:9083 May 25, 2017 9:24:12 PM org.apache.spark.launcher.OutputRedirector redirect INFO: 17/05/25 21:24:12 ERROR TSaslTransport: SASL negotiation failure May 25, 2017 9:24:12 PM org.apache.spark.launcher.OutputRedirector redirect INFO: javax.security.sasl.SaslException: GSS initiate failed [Caused by GSSException: No valid credentials provided (Mechanism level: Failed to find any Kerberos tgt)] This happens when Livy tried to connect to Hive Metastore and fails with above message. The fix is to configure Zeppelin's Livy interpreter to run in yarn-cluster mode, instead of the default yarn-client mode. After you change any interpreter configuration, you will need to restart the interpreter. Below works. livy.spark.master yarn-cluster Starting HDP 2.6.x this configuration is changed OOB to yarn-cluster.
... View more
Labels:
11-12-2016
12:32 AM
2 Kudos
This is a known issue. Livy session times out by default in 60 minutes. You can change the timeout by property livy.server.session.timeout. Upon Livy session timeout the Zeppelin's Livy interpreter needs to be restarted. This will be fixed in HDP 2.6
... View more
11-04-2016
06:41 PM
This needs 2 steps In the Zeppelin shell interpreter config, enable impersonation Ensure end user has an OS account on the node where Zeppelin is running. Thanks
... View more
10-05-2016
08:08 PM
1 Kudo
Zeppelin in HDP 2.5 does not support Spark 2. A technical preview of this will land in Hortonworks Data Cloud in next month or so.
... View more
08-17-2016
04:31 PM
@subhash parise This question seems to be about Hive but Spark is also tagged on it. Can you please explain the scenario that you are trying to enable? Thanks,
... View more