You are viewing a plain text version of this content. The canonical link for it is here.
Posted to commits@cassandra.apache.org by "Jeremy Hanna (JIRA)" <ji...@apache.org> on 2018/09/12 16:35:00 UTC
[jira] [Updated] (CASSANDRA-14736) checkHintOverload will fully
reject a write or paxosCommit when only one of the recipients has a hint
backlog
[ https://issues.apache.org/jira/browse/CASSANDRA-14736?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
Jeremy Hanna updated CASSANDRA-14736:
-------------------------------------
Summary: checkHintOverload will fully reject a write or paxosCommit when only one of the recipients has a hint backlog (was: checkHintOverloaded will fully reject a write or paxosCommit when only one of the recipients has a hint backlog)
> checkHintOverload will fully reject a write or paxosCommit when only one of the recipients has a hint backlog
> -------------------------------------------------------------------------------------------------------------
>
> Key: CASSANDRA-14736
> URL: https://issues.apache.org/jira/browse/CASSANDRA-14736
> Project: Cassandra
> Issue Type: Bug
> Components: Core
> Reporter: Benedict
> Priority: Major
>
> This seems like a fairly bad availability failure, if the failure detector does not kick in either successfully, or quickly enough. Probably we should simply not write a hint for this node? A single node can struggle for reasons besides overload, so surely we should throw an exception only when it appears to be widespread?
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
---------------------------------------------------------------------
To unsubscribe, e-mail: commits-unsubscribe@cassandra.apache.org
For additional commands, e-mail: commits-help@cassandra.apache.org