Home | About | Sematext search-lucene.com search-hadoop.com
 Search Hadoop and all its subprojects:

Switch to Threaded View
Zookeeper >> mail # dev >> Review Request 15753: Reduction of waiting time during Fast Leader Election (ZOOKEEPER-1814)


Copy link to this message
-
Re: Review Request 15753: Reduction of waiting time during Fast Leader Election (ZOOKEEPER-1814)

-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
https://reviews.apache.org/r/15753/
-----------------------------------------------------------

(Updated Dec. 13, 2013, 9:11 a.m.)
Review request for zookeeper, German Blanco and fpj.
Changes
-------

Suggested changes applied.
Bugs: ZOOKEEPER-1814
    https://issues.apache.org/jira/browse/ZOOKEEPER-1814
Repository: zookeeper
Description
-------

See Zookeeper-1814.
Diffs (updated)
-----

  ./src/docs/src/documentation/content/xdocs/zookeeperAdmin.xml 1550368
  ./src/java/main/org/apache/zookeeper/server/quorum/FastLeaderElection.java 1550368
  ./src/java/test/org/apache/zookeeper/server/quorum/QuorumPeerTestBase.java 1550368
  ./src/java/test/org/apache/zookeeper/test/FLEMaxIntervalNotificationTest.java PRE-CREATION

Diff: https://reviews.apache.org/r/15753/diff/
Testing
-------

New test case checking that the parameter maxNotificationInterval can be configurable.

The test case overwrites the receiveConnection method in QuorumCnxManager class in order to avoid any answer to Fast Leader Election. Thus, FLE is forced to retry increasing exponentially the notifications interval. This interval is measured and the test case stops when 3 consecutive measures are equal. The test case will be OK if the measure is the configured in the cfg file.
File Attachments
----------------

ZOOKEEPER-1814.patch
  https://reviews.apache.org/media/uploaded/files/2013/12/11/7e510364-6df3-486e-b338-3398960ba6b0__ZOOKEEPER-1814.patch
Thanks,

Daniel Peon