Details
-
Bug
-
Status: Closed
-
Major
-
Resolution: Resolved
-
0.7.3, 0.8.0
-
Environment 1: Zeppelin 0.7.3 via HDP 2.6.5 on Centos 7
Environment 2: Zeppelin 0.8.0 on Ubuntu 14.04 via Docker
Description
Adding BigDL version 0.7.0 via Maven Central to the default Spark interpreter seems to cause behind-the-scenes stability issues. Although no errors are given on the Interpreters page after adding the dependency, attempting to run a %spark paragraph with even trivial expressions (val x = 1) results in an indefinitely pending cell. Attempting to restart the interpreter results in nothing, and (usually) loading other notebooks no longer becomes possible. After checking both the interpreter and zeppelin server log files, I couldn't find any lines pertaining to why those paragraphs are pending indefinitely.
Attachments
Attachments
Issue Links
- is related to
-
ZEPPELIN-3856 Zeppelin add Hadoop Submarine (machine learning) interpreter
-
- Open
-
-
ZEPPELIN-3689 Shade all dependencies of zeppelin-interpreter
-
- Closed
-