[ https://issues.apache.org/jira/browse/HIVE-12330?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
Szehon Ho updated HIVE-12330: ----------------------------- Assignee: Sergio Peña (was: Szehon Ho) Sergio mentioned he might be interested. A long-term fix needs to be done for this. Right now, I am just setting qFileTests.xxx.isolateDriver= true for all the drivers as a workaround on the server. Potential ideas: 1. clean only the testdriver class files and not the java file (to force recompilation, but not remove the testdriver java file and make other test processes have no test to run). 2. disable parallelism in code for all generated testdriver files 3. rework the generated testdriver file such that they read an environment variable (probably the cleanest, needs investigation) instead of hardcode the path > Fix precommit Spark test part2 > ------------------------------ > > Key: HIVE-12330 > URL: https://issues.apache.org/jira/browse/HIVE-12330 > Project: Hive > Issue Type: Bug > Components: Testing Infrastructure > Reporter: Szehon Ho > Assignee: Sergio Peña > Attachments: HIVE-12229.3-spark.patch > > > Regression because of HIVE-11489 -- This message was sent by Atlassian JIRA (v6.3.4#6332)