question Re: Configuration Steps for Central Repository of Zeppelin Notebooks in a Multi-Node Environment in Archives of Support Questions (Read Only)

Configuration Steps for Central Repository of Zeppelin Notebooks in a Multi-Node Environment

bhanu_pamu — Wed, 23 May 2018 17:10:16 GMT

I have installed Zeppelin Notebooks on two nodes in a Ambari Managed HDP Cluster using the link. Both the links are up and running. User authentication is enabled via LDAP users.

When a user is creating a notebook on Node 1, the notebook is not visible when he logs in from URL-of node 2.

zeppelin.notebook.storage=org.apache.zeppelin.notebook.repo.FileSystemNotebookRepo is enabled.

The version of Zeppelin is 0.7.3

Is it possible to use - HDFS file system location as the repository path for Zeppelin notebooks and where do I configure the path for Zeppelin, so that both nodes point to the same repo.

Thanks.

Re: Configuration Steps for Central Repository of Zeppelin Notebooks in a Multi-Node Environment

anarasimham — Wed, 23 May 2018 20:49:06 GMT

@Bhanu Pamu

Please refer to the following section of our documentation to enable HDFS storage for Zeppelin HA:

https://docs.hortonworks.com/HDPDocuments/HDP2/HDP-2.6.3/bk_zeppelin-component-guide/content/ch_zeppelin_upgrade_hdfs_storage.htm

Re: Configuration Steps for Central Repository of Zeppelin Notebooks in a Multi-Node Environment

bhanu_pamu — Thu, 24 May 2018 17:30:21 GMT

I followed the above steps. The configuration in Ambari --> Zeppelin Notebook--> Configs --> Advanced zeppelin-config

zeppelin.config.fs.dir=conf
zeppelin.notebook.dir=notebook
zeppelin.interpreter.dir=interpreter
zeppelin.notebook.storage=org.apache.zeppelin.notebook.repo.FileSystemNotebookRepo

Yet while I login to Zeppelin Notebook to either of the nodes - the Notebook Repos tab shows the following value

Notebook Path

/usr/hdp/current/zeppelin-server/notebook/

So whenever I create a notebook, it goes and saved in local zeppelin server directory.

I tried changing the path of Network Repos from Zeppelin login, however when I restart the services, it goes back to default.

I tried changing the value of ZEPPELIN_NOTEBOOK_DIR in zeppelin_env_content using Ambari, however the after service restart Zeppelin doesn't work properly. It gives 503 server error or the notebooks repo is empty, no old notebooks seen. I have copied the notebooks from /usr/hdp/current/zeppelin-server/notebook/local to hdfs file system /user/zeppelin/notebook, still no gain.

Somehow the configuration doesn't work in HA mode. The notebook files are not written to HDFS file system.

Any further suggestions?

Re: Configuration Steps for Central Repository of Zeppelin Notebooks in a Multi-Node Environment

bhanu_pamu — Thu, 24 May 2018 18:23:55 GMT

Followed the steps mentioned in this link - and it worked.

https://community.hortonworks.com/questions/149389/zeppelin-does-not-write-its-notes-in-hdfs-on-hdp-2.html

Re: Configuration Steps for Central Repository of Zeppelin Notebooks in a Multi-Node Environment

layne_sadler — Sun, 21 Jul 2019 13:02:48 GMT

same issue. It seems like Notebook Path is not getting persisted in the config anywhere.

Re: Configuration Steps for Central Repository of Zeppelin Notebooks in a Multi-Node Environment

ask_bill_brooks — Mon, 22 Jul 2019 06:36:05 GMT

The above question and the entire response thread below were originally posted in the Community Help track. On Sun Jul 21 23:35 UTC 2019, a member of the HCC moderation staff moved it to the Data Science & Advanced Analytics track. The Community Help Track is intended for questions about using the HCC site itself.