You are viewing a plain text version of this content. The canonical link for it is here.

Posted to oak-issues@jackrabbit.apache.org by "Michael Dürig (JIRA)" <ji...@apache.org> on 2016/11/15 15:20:58 UTC

[jira] [Updated] (OAK-5111) Change default size of the node deduplication cache

     [ https://issues.apache.org/jira/browse/OAK-5111?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]

Michael Dürig updated OAK-5111:
-------------------------------
    Attachment: node-dedup-cache.xlsx

Excel sheet showing cleanup and compaction timing and efficacy for different cache sizes on a longevity run of 13 days: [^node-dedup-cache.xlsx]. 

There is virtually no different between cache sizes 1, 2 and 4M. 8M seems worse, most likely due to memory pressure. < 200k is also worse as most likely there are too many cache misses. I would thus change the default cache size to 1M. 

> Change default size of the node deduplication cache
> ---------------------------------------------------
>
>                 Key: OAK-5111
>                 URL: https://issues.apache.org/jira/browse/OAK-5111
>             Project: Jackrabbit Oak
>          Issue Type: Improvement
>          Components: segment-tar
>            Reporter: Michael Dürig
>            Assignee: Michael Dürig
>              Labels: compaction, gc, perfomance
>             Fix For: 1.6, 1.5.14
>
>         Attachments: node-dedup-cache.xlsx
>
>
> The current default of the node deduplication cache is 8M. We should consider changing this to a smaller value still resulting in effective compactions. 



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