You are viewing a plain text version of this content. The canonical link for it is here.
Posted to notifications@couchdb.apache.org by "Bala Kolla (JIRA)" <ji...@apache.org> on 2016/09/28 21:09:21 UTC

[jira] [Resolved] (COUCHDB-3150) Update all shards with stale=update_after

     [ https://issues.apache.org/jira/browse/COUCHDB-3150?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]

Bala Kolla resolved COUCHDB-3150.
---------------------------------
    Resolution: Fixed

> Update all shards with stale=update_after
> -----------------------------------------
>
>                 Key: COUCHDB-3150
>                 URL: https://issues.apache.org/jira/browse/COUCHDB-3150
>             Project: CouchDB
>          Issue Type: Story
>          Components: Database Core
>            Reporter: Bala Kolla
>
> Stale=update_after should update shards that won't participate in the response due to not being a ushard.
> Discussed with Paul Davis and he suggested the following approach.
> For this ticket we should just need to add a fabric_view:maybe_update_others/3 that we call just after [1,2] with fabric_view:maybe_update_others(DbName, Shards, Args). And then if stable=true and update=(true or lazy) we do mem3:shards(DbName) -- ShardsInvolved, and throw a message at them to trigger an update.
> [1] https://github.com/apache/couchdb-fabric/blob/master/src/fabric_view_map.erl#L27
> [2] https://github.com/apache/couchdb-fabric/blob/master/src/fabric_view_reduce.erl#L28



--
This message was sent by Atlassian JIRA
(v6.3.4#6332)