[
https://issues.apache.org/jira/browse/CASSANDRA-1350?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12897345#action_12897345
]
Brandon Williams commented on CASSANDRA-1350:
---------------------------------------------
The latencies being cleared is the most important part. At the next interval,
if no more latency data is received for that node, updateScores() going to
calculate a score for an empty deque, which means the score() function is just
to going to do a size check and return 0, which is pretty cheap, so I'm not
sure it's worth the extra complexity of notifying the snitch about
decommissioned nodes.
> DynamicEndpointSnitch is defeated by the caching done in Strategy
> -----------------------------------------------------------------
>
> Key: CASSANDRA-1350
> URL: https://issues.apache.org/jira/browse/CASSANDRA-1350
> Project: Cassandra
> Issue Type: Bug
> Components: Core
> Affects Versions: 0.7 beta 1
> Reporter: Jonathan Ellis
> Assignee: Gary Dusbabek
> Fix For: 0.7 beta 1
>
> Attachments: 0001-move-endpoint-cache-from-ARS-to-AES.patch
>
>
> can we move the caching into AbstractEndpointSnitch instead?
> also: AES.register appears to never be called.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
