You are viewing a plain text version of this content. The canonical link for it is here.
Posted to issues@ratis.apache.org by "runzhiwang (Jira)" <ji...@apache.org> on 2020/08/02 05:34:00 UTC
[jira] [Updated] (RATIS-993) Preventing disruptions when a server
rejoins the cluster
[ https://issues.apache.org/jira/browse/RATIS-993?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
runzhiwang updated RATIS-993:
-----------------------------
Summary: Preventing disruptions when a server rejoins the cluster (was: Pre vote before request vote)
> Preventing disruptions when a server rejoins the cluster
> --------------------------------------------------------
>
> Key: RATIS-993
> URL: https://issues.apache.org/jira/browse/RATIS-993
> Project: Ratis
> Issue Type: New Feature
> Reporter: runzhiwang
> Assignee: runzhiwang
> Priority: Major
>
> For example, leader is s1, and s2, s3 are the followers. If there is something wrong with s3's network, and s3 lose connection with leader in a short time interval, s3 will trigger election. when s3's network recovery, leader will step down which is unnecessary. So I want to add pre vote before real vote.
> 1. if s3 want to request vote, it first send a pre vote request.
> 2. when s1 and s2 receive pre vote request of s3, they check their electionTimeout to decide whether should trigger leader election.
> 3. If s1 and s2 refuse to trigger a new leader election, s3 still keep as follower.
--
This message was sent by Atlassian Jira
(v8.3.4#803005)