You are viewing a plain text version of this content. The canonical link for it is here.
Posted to issues@kudu.apache.org by "Todd Lipcon (JIRA)" <ji...@apache.org> on 2018/01/19 02:01:00 UTC
[jira] [Commented] (KUDU-1720) Expose ability to appoint a replica
as leader via RPC
[ https://issues.apache.org/jira/browse/KUDU-1720?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16331611#comment-16331611 ]
Todd Lipcon commented on KUDU-1720:
-----------------------------------
[~mpercy] [~wdberkeley] do you guys think this is still necessary?I seem to recall Will had written instructions for recovering from the kind of scenario mentioned here using the existing unsafe_change_config tool.
> Expose ability to appoint a replica as leader via RPC
> -----------------------------------------------------
>
> Key: KUDU-1720
> URL: https://issues.apache.org/jira/browse/KUDU-1720
> Project: Kudu
> Issue Type: Bug
> Components: consensus
> Reporter: Mike Percy
> Priority: Major
>
> Today we have an internal API call in RaftConsensus called EmulateElection() that causes a node to increment its term and begin acting as Raft leader without initiating an election. Exposing this would allow us to use this in emergency scenarios, where all replicas except one are offline and we need to force a node to become leader in order to then force an unsafe configuration change to be serialized to the WAL.
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)