Does the HiveContext object expire in Zeppelin after a period of time?

[#1]

I am seeing behaviour where I open a notebook and run a paragraph script in Zeppelin and the script executes and returns the result as expected. Then I leave the notebook open for a few hours - come back and attempt to re-run the same paragraph and get the error: "object 'hiveContext' not found". The only way to resolve this is by restarting the Spark interpreter. Is this expected behaviour? Is there a timeout on the hiveContext perhaps?

Any insight much appreciated

[#2]

..also is there a way in SparkR to check first if the sparkcontext and hivecontexts are running?

[#3]

Hi @mike harding

What version of Zeppelin are you using?

[#4]

Its 0.6.0 > https://zeppelin.apache.org/download.html

[#5]

@mike harding -> are you running Zeppelin (within HDP 2.5 Sandbox) or running Zeppelin standalone?

[#6]

running on HDP2.4

[#7]

Are you using Zeppelin directly with Spark interpreter or with Livy interpreter?

In Z > Livy > Spark scenario, the Livy session is recycled some time of inactivity. I will report back if default timeout and how to configure it.

To come out of this behavior, you will need to restart Livy interpreter.

[#8 · Accepted solution]

Livy sessions are recycled after an hour of session inactivity. This timeout is configured with livy.server.session.timeout