cassandra-commits mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Stefan Podkowinski (JIRA)" <j...@apache.org>
Subject [jira] [Commented] (CASSANDRA-12870) Calculate pending ranges for identical KS settings just once
Date Mon, 20 Mar 2017 14:48:41 GMT

    [ https://issues.apache.org/jira/browse/CASSANDRA-12870?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15932756#comment-15932756
] 

Stefan Podkowinski commented on CASSANDRA-12870:
------------------------------------------------

Elements from the keyspace list will be removed by keyspaces.iterator().remove(). That's why
the list is copied before as I don't want to remove elements from an "unowned" collection.
LinkedList was chosen for performance reasons, although totally insignificant at such low
cardinality. The LHS declaration should be changed to List<String> to make the code
style coherent, yes. 

> Calculate pending ranges for identical KS settings just once
> ------------------------------------------------------------
>
>                 Key: CASSANDRA-12870
>                 URL: https://issues.apache.org/jira/browse/CASSANDRA-12870
>             Project: Cassandra
>          Issue Type: Improvement
>          Components: Core
>            Reporter: Stefan Podkowinski
>            Assignee: Stefan Podkowinski
>             Fix For: 4.0
>
>         Attachments: 12870-trunk.patch
>
>
> The {{TokenMetadata.calculatePendingRanges()}} operation can be very expensive and already
has been subject to micro optimization in CASSANDRA-9258. Instead of further optimizing the
calculation itself, I'd like to prevent it from being executed as often by only calling it
just once for all keyspaces sharing identical replication settings. 



--
This message was sent by Atlassian JIRA
(v6.3.15#6346)

Mime
View raw message