[
https://issues.apache.org/jira/browse/FLUME-2298?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13877311#comment-13877311
]
Ashish Paliwal commented on FLUME-2298:
---------------------------------------
Diagram: On the left it would be Flume Source before put proxy?
IMHO, DC channel implementation should have Avro Source/Sink. The Master/Slave
is just a state, else we end up with Master DC and Slave DC implementation.
Ordering in Slave - Slave won't be shared across multiple masters, it should be
replica of Master Queue. Not able to understand the cases where ordering shall
help. May be missing something.
Slave Channel needs to have Sink settings from master, if it has to drain the
events. If master goes down, like node failure, the sink/Take Proxy running
there also goes down.
So the only thing that can happen is draining the Channel on Slave, not sure if
having Take proxy would help.
BTW, Would it help if we call it Replicated Channel instead of DC?
> Distributed Channel
> -------------------
>
> Key: FLUME-2298
> URL: https://issues.apache.org/jira/browse/FLUME-2298
> Project: Flume
> Issue Type: New Feature
> Components: Channel
> Reporter: Ted Malaska
> Assignee: Ted Malaska
> Attachments: FlumeDistributedChannelDesign.0.1.pdf,
> FlumeDistributedChannelDesign.0.2.pdf
>
>
> This channel will allow for events to be persisted with a plugable method on
> more then one agent or node.
> The goal is to gain the following benefits:
> 1. Events will continue to flow to sinks with out loss or with out large
> delay even in the case of node failure.
> 2. Protect against loss in the case of a complete single node failure
--
This message was sent by Atlassian JIRA
(v6.1.5#6160)
