----------------------------------------------------------- This is an automatically generated e-mail. To reply, visit: https://reviews.apache.org/r/27292/#review58948 -----------------------------------------------------------
shims/0.23/src/main/java/org/apache/hadoop/hive/shims/Hadoop23Shims.java <https://reviews.apache.org/r/27292/#comment100156> - That's correct, we'll reload the config for each new session. - Calling setReloadListener() is the only way to get back the full AllocationConfiguration (including the yarn config and fair-scheduler.xml). Without the callback, we can't access those details from the calling method. Does that answer the question ? - yes, we don't really need to protect the config. The AtomicReference it just a workaround to retriev the refernced object to a nested block - Prasad Mujumdar On Oct. 28, 2014, 5:48 p.m., Prasad Mujumdar wrote: > > ----------------------------------------------------------- > This is an automatically generated e-mail. To reply, visit: > https://reviews.apache.org/r/27292/ > ----------------------------------------------------------- > > (Updated Oct. 28, 2014, 5:48 p.m.) > > > Review request for hive, Brock Noland and Mohit Sabharwal. > > > Bugs: HIVE-8634 > https://issues.apache.org/jira/browse/HIVE-8634 > > > Repository: hive-git > > > Description > ------- > > The fair scheduler queue refresh in HiveServer2 (for non-impersonation mode), > doesn't handle the primary/secondary queue mappings correctly. It's not > reading primary and secondary rules from the scheduler rule file. > The patch involves refreshing the queue policy allocation to read the > fair-scheduler.xml file correctly. > > > Diffs > ----- > > data/conf/fair-scheduler-test.xml PRE-CREATION > > itests/hive-unit-hadoop2/src/test/java/org/apache/hive/jdbc/TestSchedulerQueue.java > 79878ba > shims/0.23/src/main/java/org/apache/hadoop/hive/shims/Hadoop23Shims.java > 6125714 > > Diff: https://reviews.apache.org/r/27292/diff/ > > > Testing > ------- > > Added test cases to for various queue allocation rules. > > > Thanks, > > Prasad Mujumdar > >
