[jira] Updated: (COUCHDB-351) intra-document checkpointing during replication

["Paul Joseph Davis (JIRA)" <ji...@apache.org>]

     [ https://issues.apache.org/jira/browse/COUCHDB-351?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]

Paul Joseph Davis updated COUCHDB-351:
--------------------------------------

    Skill Level: Committers Level (Medium to Hard)

> intra-document checkpointing during replication
> -----------------------------------------------
>
>                 Key: COUCHDB-351
>                 URL: https://issues.apache.org/jira/browse/COUCHDB-351
>             Project: CouchDB
>          Issue Type: New Feature
>          Components: Database Core
>            Reporter: Adam Kocoloski
>            Priority: Minor
>
> Ben Cohen noted that Couch will fail to make progress if a link is never up long enough to transfer a specific large document or attachment:
> http://mail-archives.apache.org/mod_mbox/couchdb-user/200905.mbox/%3cA192849F-B7EE-48A2-B010-D63D171FC034@gmail.com%3e
> If we want to do intra-document checkpointing we'll need:
> 1) support for the Range header in GET requests
> 2) a place to store partially-transferred docs/attachments when a replication fails -- possibly as attachments to the replicator _local doc.  On the other hand, this feature is not really replication-specific, so we'd probably want some sort of standalone "upload cache" that saves partial transfers.
> 3) for push replication, an extension to the _missing_revs API that can describe partial transfers existing on the target

-- 
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.