You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@helix.apache.org by "Hudson (JIRA)" <ji...@apache.org> on 2018/07/09 19:02:00 UTC
[jira] [Commented] (HELIX-711) Allow downward state transition
during recovery and add recovery threshold
[ https://issues.apache.org/jira/browse/HELIX-711?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16537413#comment-16537413 ]
Hudson commented on HELIX-711:
------------------------------
FAILURE: Integrated in Jenkins build helix #1480 (See [https://builds.apache.org/job/helix/1480/])
[HELIX-711] Allow downward state transition during recovery and add (narendly: rev 37f3d4c8dadd7cadeebad4fb41e2d4b1c38601fa)
* (edit) helix-core/src/test/java/org/apache/helix/controller/stages/TestIntermediateStateCalcStage.java
* (add) helix-core/src/test/java/org/apache/helix/controller/stages/TestRecoveryLoadBalance.java
* (edit) helix-core/src/main/java/org/apache/helix/controller/stages/IntermediateStateCalcStage.java
* (add) helix-core/src/test/resources/TestRecoveryLoadBalance.MasterSlave.json
* (add) helix-core/src/test/resources/TestRecoveryLoadBalance.OnlineOffline.json
* (edit) helix-core/src/main/java/org/apache/helix/api/config/StateTransitionThrottleConfig.java
* (edit) helix-core/src/main/java/org/apache/helix/model/ClusterConfig.java
> Allow downward state transition during recovery and add recovery threshold
> --------------------------------------------------------------------------
>
> Key: HELIX-711
> URL: https://issues.apache.org/jira/browse/HELIX-711
> Project: Apache Helix
> Issue Type: Improvement
> Reporter: Hunter L
> Priority: Major
>
> # Previously, a single partition requiring recovery balance would block all types of load-balance. This commit allows all downward state transitions (load balance) to happen even when recovery balance is happening in the same cycle.
> # As for non-downward state transitions load-balance, a parameter, ErrorOrRecoveryPartitionThresholdForLoadBalance, was added to ClusterConfig. If the number of partitions requiring recovery is lower than the threshold, non-downward load-balance will take place in the same cycle as recovery balance; otherwise, non-downward load-balance will not take place in the same cycle.
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)