jackrabbit-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Jukka Zitting (JIRA)" <j...@apache.org>
Subject [jira] Commented: (JCR-2770) Initial size of ConcurrentCache depends on number of segments (available processors)
Date Wed, 06 Oct 2010 13:22:33 GMT

    [ https://issues.apache.org/jira/browse/JCR-2770?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12918509#action_12918509
] 

Jukka Zitting commented on JCR-2770:
------------------------------------

> What is the reason that you set the initial size to 1024?

The three-argument constructor (that's needed to enable LRU behaviour) requires some value
there, and since there's no public constant for the default value I simply selected something
that seemed reasonable. I chose a high value since I expected most practical cases to in any
case use pretty large caches, and since environments with lots of processors would typically
also have lots of available memory. I of course forgot to consider the fixed memory limits
we specify for our test runs.

Essentially I just picked the value from my proverbial hat, so it's fine to change it to something
different, especially if doing so helps prevent problems!

> Initial size of ConcurrentCache depends on number of segments (available processors)
> ------------------------------------------------------------------------------------
>
>                 Key: JCR-2770
>                 URL: https://issues.apache.org/jira/browse/JCR-2770
>             Project: Jackrabbit Content Repository
>          Issue Type: Improvement
>          Components: jackrabbit-core
>            Reporter: Marcel Reutegger
>            Priority: Minor
>             Fix For: 2.2.0
>
>         Attachments: JCR-2770.patch
>
>
> This causes a build failure on my machine. Tests run into an OOME because the initial
memory footprint of a ConcurrentCache on my machine is 8k. Many of the tests keep references
to some kind of repository objects (node, session, x-manager), which means ConcurrentCache
instances  cannot be garbage collected immediately after a test run.
> I think the overall initial size of the cache should be independent of the number of
segments. See proposed patch.

-- 
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.


Mime
View raw message