[
https://issues.apache.org/jira/browse/COUCHDB-2842?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14947088#comment-14947088
]
Robert Newson commented on COUCHDB-2842:
----------------------------------------
This proved tricky as replicator docs for the same logical job can be run on
different nodes (doc id being part of the 'owner' function).
Leaving this open until we come up with a smart way to stop duplicates.
> Fix duplicate _replicator docs
> ------------------------------
>
> Key: COUCHDB-2842
> URL: https://issues.apache.org/jira/browse/COUCHDB-2842
> Project: CouchDB
> Issue Type: Improvement
> Security Level: public(Regular issues)
> Reporter: Robert Newson
>
> Duplicate jobs transition directly to completed state
> _replicate is idempotent but _replicator is not, one can create any
> number of documents with the same content. The burden of retrying all
> the duplicates can be excessive.
> A detected duplicate will now transition immediately to completed, as
> we know there's a currently running job for the exact same task.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
