[
https://issues.apache.org/jira/browse/SOLR-2808?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13120151#comment-13120151
]
Yonik Seeley commented on SOLR-2808:
------------------------------------
When a node comes up, it starts in "recovering" state (actually, this is
probably a per-shard state, but for now, one core == one shard).
While in the recovering state, updates should still be sent to the node, but
these updates should be buffered (most likely using transaction logs) to be
applied later.
The node should ask peer(s) (or the leader?) for updates it missed. If peers
can't provide those updates (because the transaction logs only go back so far),
then the node needs to initiate a full index copy first (probably re-using java
replication code).
After the new index is in place, updates from peers have been applied, and
buffered updates have been applied, the state can move to "active".
> Node Recovery
> -------------
>
> Key: SOLR-2808
> URL: https://issues.apache.org/jira/browse/SOLR-2808
> Project: Solr
> Issue Type: Sub-task
> Components: SolrCloud
> Reporter: Yonik Seeley
> Fix For: 4.0
>
>
> Node recovery encompasses everything involved in getting a node to the
> "active" state after coming (back) up.
--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators:
https://issues.apache.org/jira/secure/ContactAdministrators!default.jspa
For more information on JIRA, see: http://www.atlassian.com/software/jira
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
