[
https://issues.apache.org/jira/browse/COUCHDB-2516?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14249037#comment-14249037
]
Paul Joseph Davis commented on COUCHDB-2516:
--------------------------------------------
[~kxepal] There are two current methods. One as Ilya points out is if a
document is modified during compaction, when compaction runs its catchup phase,
the document's attachments will be recopied.
The second method is if a document has multiple revisions that share an
identical attachment. Each leaf of the revision tree will copy its attachments
over indepentently as well which copies the attachments unnecessarily.
Ilya's work on this ticket also goes as far as deduplicating attachments
between documents as well which is a nice added bonus for possible space
savings.
> Deduplicate attachements on compaction
> --------------------------------------
>
> Key: COUCHDB-2516
> URL: https://issues.apache.org/jira/browse/COUCHDB-2516
> Project: CouchDB
> Issue Type: New Feature
> Security Level: public(Regular issues)
> Components: Database Core
> Reporter: ILYA
>
> Currently we copy all attachments to .compact file even if they've been
> copied already on compacting previous revision of the document.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
