[
https://issues.apache.org/jira/browse/OAK-8673?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16973185#comment-16973185
]
Thomas Mueller commented on OAK-8673:
-------------------------------------
> beyond the task at hand to re-evaluate if the current value of
> eager-cache-size is sufficient
Well you don't want to expand the cache size if there is a risk of running out
of memory... But given the next statement I'm not sure if there really is such
a risk...
> even for the lazy-evaluation a cache is populated (in fact there are even 2
> maps in that case), so depending on the distribution of permission entries
> and the access pattern (read/writing), the lazy cache might even consume more
> memory than the eager-cache...
But, why are benchmark results so bad the eager cache is disabled (size set to
0)?
> Determine and possibly adjust size of eagerCacheSize
> ----------------------------------------------------
>
> Key: OAK-8673
> URL: https://issues.apache.org/jira/browse/OAK-8673
> Project: Jackrabbit Oak
> Issue Type: Technical task
> Components: core, security
> Reporter: Angela Schreiber
> Assignee: Angela Schreiber
> Priority: Major
>
> The initial results of the {{EagerCacheSizeTest}} seem to indicate that we
> almost never benefit from the lazy permission evaluation (compared to reading
> all permission entries right away). From my understanding of the results the
> only exception are those cases where only very few items are being accessed
> (e.g. reading 100 items).
> However, I am not totally sure if this is not a artifact of the random-read.
> I therefore started extending the benchmark with an option to re-read a
> randomly picked item more that once, which according to some analysis done
> quite some time ago is a common scenario specially when using Oak in
> combination with Apache Sling.
> Benchmarks with 10-times re-reading the same random item:
> As I would have expected it seems that the negative impact of lazy-loading is
> somewhat reduced, as the re-reading will hit the cache populated while
> reading.
> Result are attached to OAK-8662 (possibly more to come).
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
