[
https://issues.apache.org/jira/browse/TOREE-516?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17071306#comment-17071306
]
Luciano Resende commented on TOREE-516:
---------------------------------------
While Toree kernel can be submitted in yarn cluster mode, vanilla Jupyter
Notebooks expect kernels as local processes and cannot discover where in the
YARN cluster the actual kernel launched. The proper way to handle this is to
have Jupyter Notebook -> Jupyter Enterprise Gateway -> Toree Kernel, and this
is described in
https://jupyter-enterprise-gateway.readthedocs.io/en/latest/kernel-yarn-cluster-mode.html
> Kerberos error while working with Toree Kernel
> ----------------------------------------------
>
> Key: TOREE-516
> URL: https://issues.apache.org/jira/browse/TOREE-516
> Project: TOREE
> Issue Type: Bug
> Components: Kernel
> Affects Versions: 0.3.0
> Environment: RHEL 7.7, HDP 3 , Spark 2.3.2
> Reporter: Bharath
> Priority: Critical
> Labels: security
>
> We have a Kerberied HDP 3.1.0 cluster, configured Jupyter/Toree - It throws
> Keberos error, we have a valid kerberos ticket in place. Here is the
> kernel.json
> {
> "argv": [
> "/opt/jupyterhub/share/jupyter/kernels/apache_toree_scala/bin/run.sh",
> "--profile",
> "\{connection_file}"
> ],
> "env": {
> "DEFAULT_INTERPRETER": "Scala",
> "TOREE_SPARK_OPTS": "--master yarn --deploy-mode cluster",
> "TOREE_OPTS": "",
> "HADOOP_CONF_DIR": "/etc/hadoop/conf",
> "SPARK_HOME": "/usr/hdp/3.1.0.0-78/spark2",
> "PYTHONPATH":
> "/opt/jupyterhub/bin/python:/opt/anaconda3/bin/python:/usr/hdp/3.1.0.0-78/spark2/python/lib/py4j-0.10.7-src.zip:/usr/hdp/3.1.0.0-78/spark2/python:/usr/hdp/3.1.0.0-78/spark2/python/lib",
> "PYTHON_EXEC": "python",
> "HBASE_CONF_DIR": "/etc/hbase/conf",
> "HBASE_HOME": "/usr/hdp/current/hbase-client",
> "JAVA_HOME": "/usr/lib/jvm/java-1.8.0-openjdk-1.8.0.232.b09-0.el7_7.x86_64"
> },
> "display_name": "Toree - Scala-kernel",
> "language": "scala",
> "interrupt_mode": "signal",
> "metadata": {}
> }
> Error Message:
> 20/03/26 18:32:28 INFO RpcRetryingCallerImpl: Call exception, tries=8,
> retries=36, started=18874 ms ago, cancelled=false, msg=Call to
> compute-4.datalake.ntt/xxx.xxx.xx.xxx:16020 failed on local exception:
> javax.security.sasl.SaslException: GSS initiate failed [Caused by
> GSSException: No valid credentials provided (Mechanism level: Failed to find
> any Kerberos tgt)], details=row 'name_data:exchg_yearly,,99999999999999' on
> table 'hbase:meta' at region=hbase:meta,,1.1588230740,
> hostname=compute-4.datalake.ntt
> With above configuration, kernel dies sometimes.
> Please suggest if the way we are passing configs in Kernel.Json is the right
> way ?
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
