[
https://issues.apache.org/jira/browse/COUCHDB-2492?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14736612#comment-14736612
]
Robert Newson commented on COUCHDB-2492:
----------------------------------------
I don't think we should do this on the clustered interface. Like BigCouch, we
should continue to direct admins to use the backend, per-node interface.
Compacting all shards of a db or view at once on all nodes is a significantly
bigger event than it is in couchdb before 2.0. While we do have the io
regulator, there will still be an impact.
At least for 2.0, I would rather we didn't do this work. The compaction daemon
should be used instead. Compaction is a node-local issue imo.
> Fauxton: Compact Databases & Compact View does not work on 2.0
> --------------------------------------------------------------
>
> Key: COUCHDB-2492
> URL: https://issues.apache.org/jira/browse/COUCHDB-2492
> Project: CouchDB
> Issue Type: Bug
> Security Level: public(Regular issues)
> Components: BigCouch, Fauxton
> Reporter: Robert Kowalski
> Priority: Blocker
>
> Select a database, select "Compact & Clean", press the button "compact":
> You'll get:
> Error: Compaction must be triggered on a per-shard basis in CouchDB
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
