[
https://issues.apache.org/jira/browse/ZOOKEEPER-2136?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14357308#comment-14357308
]
Flavio Junqueira commented on ZOOKEEPER-2136:
---------------------------------------------
The options I'm seeing are:
# Make sync a quorum operation, which is what this jira originally proposed
# Create new calls for quorum reads
# Make no changes and use write operations to sync
>From all these, the one that seems most reasonable is the first one because it
>requires no API changes and actually cleans up the code. Having quorum reads
>requires extending the API (in a BC way), which doesn't feel strictly
>necessary.
> Sync() should get quorum acks.
> ------------------------------
>
> Key: ZOOKEEPER-2136
> URL: https://issues.apache.org/jira/browse/ZOOKEEPER-2136
> Project: ZooKeeper
> Issue Type: Sub-task
> Reporter: Hongchao Deng
> Assignee: Hongchao Deng
> Fix For: 3.6.0
>
> Attachments: ZOOKEEPER-2000.patch
>
>
> Currently if the sync packet goes to leader it doesn't get quorum acks. This
> is a problem during reconfig and leader changes. testPortChange() flaky
> failure is caused by such case.
> I proposed to change sync() semantics to require quorum acks in any case.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
