cassandra-commits mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Chris Burroughs (JIRA)" <j...@apache.org>
Subject [jira] [Commented] (CASSANDRA-9872) only check KeyCache when it is enabled
Date Mon, 17 Aug 2015 18:10:45 GMT

    [ https://issues.apache.org/jira/browse/CASSANDRA-9872?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14699957#comment-14699957
] 

Chris Burroughs commented on CASSANDRA-9872:
--------------------------------------------

I *think* this is correct (and simpler!) with all of the 3.0 branch changes.  I looked into
adding unit tests to `KeyCacheTest` but it's pretty end to end and I didn't see any thing
existing tests that call `getCachedPosition` directly.

> only check KeyCache when it is enabled
> --------------------------------------
>
>                 Key: CASSANDRA-9872
>                 URL: https://issues.apache.org/jira/browse/CASSANDRA-9872
>             Project: Cassandra
>          Issue Type: Improvement
>          Components: Core
>            Reporter: Chris Burroughs
>            Assignee: Chris Burroughs
>              Labels: cache, metrics
>         Attachments: j9872-2.0-v1.txt, j9872-3.0-v1.txt
>
>
> If the KeyCache exists (because at least one column family is using it) we currenlty
check the key cache even for requests to column families where the key cache is disabled.
 I think it would be better to only check the cache if entries *could* be there.
>  * This will align the key cache with how the row cache behaves.
>  * This makes the key cache metrics much more useful.  For example, 'requests' becomes
'requests to things that could be in the key cache' and not just 'total requests'.
>  * This migh be a micro-optimization saving a few metric update.



--
This message was sent by Atlassian JIRA
(v6.3.4#6332)

Mime
View raw message