[
https://issues.apache.org/jira/browse/CASSANDRA-1103?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12868808#action_12868808
]
Jeremy Hanna edited comment on CASSANDRA-1103 at 5/18/10 4:07 PM:
------------------------------------------------------------------
Looks good, makes it more readable with the refactor.
Two comments:
1) DSS: in loadEndpoints, for dcTokens, could a Collection be used that is
naturally sorted so it doesn't have to iterate again through to sort them?
2) should the comment on line 156 really be before line 160 instead where the
loop is
was (Author: jeromatron):
Looks good, makes it more readable with the refactor.
Two comments:
1) DSS: in loadEndpoints, for dcTokens, could a Collection be used that is
naturally sorted so it doesn't have to iteration through them to sort them?
2) should the comment on line 156 really be before line 160 instead where the
loop is
> DSS rack-awareness doesn't really work
> --------------------------------------
>
> Key: CASSANDRA-1103
> URL: https://issues.apache.org/jira/browse/CASSANDRA-1103
> Project: Cassandra
> Issue Type: Bug
> Components: Core
> Affects Versions: 0.6
> Reporter: Jonathan Ellis
> Assignee: Jonathan Ellis
> Fix For: 0.7
>
> Attachments:
> 0001-clean-up-DSS-inline-getNaturalEndpointsInternal-remove.txt,
> 0002-simplify-out-doneDataCenterIter-variable.txt,
> 0003-fix-DSS-to-generate-unique-racks-when-possibleb.txt
>
>
> CASSANDRA-952 fixed most of the DSS issues but the attempted placement of
> machines on different racks w/in each DC is poor (comparing each node only to
> the rack of the 1st replica in that DC rather than all).
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
