[ZEPPELIN-3205] restarting interpreter setting in a notebook abort running jobs of other notebooks

• Type: Bug
• Status: Open
• Priority: Major
• Resolution: Unresolved
• Affects Version/s: 0.7.3
• Assignee: Unassigned
• Reporter: dungnguyen
• Created: 06/Feb/18 23:17
• Updated: 07/Feb/18 05:24

Description

I'm aware that there is resolved issues

https://issues.apache.org/jira/browse/ZEPPELIN-1770 

But it's pretty simple to reproduce, I can configure spark or python interpreters in per-note isolated mode, and start a long running job in 2 notebooks #1 and #2. If I restart spark or python (depends on the type of running job) interpreter in notebook #1, the job in notebook #2 is aborted. It is worse for pyspark since not just the job is aborted, the pyspark python process of notebook #2 is also killed, but notebook #2 will be hanging afterward, the only way to fix is to restart notebook #2

I also found a related issue for python interpreter

https://issues.apache.org/jira/browse/ZEPPELIN-3171

Issue Links

• duplicates: ZEPPELIN-3171 Restart of interpreter in note also aborts running interpreter in another note