[
https://issues.apache.org/jira/browse/CASSANDRA-517?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Jonathan Ellis updated CASSANDRA-517:
-------------------------------------
Attachment: 0003-force-ranges-to-be-increasing-to-avoid-bug-in-midpoint.txt
0002-prefer-bootstrapping-into-ring-sections-that-are-not-a.txt
0001-CASSANDRA-517-make-AbstractReplicationStrategy-range-m.txt
> have bootstrap token-selector consider predicted load
> -----------------------------------------------------
>
> Key: CASSANDRA-517
> URL: https://issues.apache.org/jira/browse/CASSANDRA-517
> Project: Cassandra
> Issue Type: New Feature
> Components: Core
> Reporter: Jonathan Ellis
> Assignee: Jonathan Ellis
> Fix For: 0.5
>
> Attachments:
> 0001-CASSANDRA-517-make-AbstractReplicationStrategy-range-m.txt,
> 0002-prefer-bootstrapping-into-ring-sections-that-are-not-a.txt,
> 0003-force-ranges-to-be-increasing-to-avoid-bug-in-midpoint.txt
>
>
> continued from CASSANDRA-483: an autobootstrapped node will pick the
> most-loaded node to get a token from. but if we autobootstrap multiple
> nodes, we don't want all of them picking tokens right next to each other in
> the same ring segment. additional bootstrapping nodes need to take this into
> consideration, ideally without needing an error-prone "rollback" of load
> guessing if one of the bootstraps fails.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
