[
https://issues.apache.org/jira/browse/CAMEL-8994?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14639139#comment-14639139
]
Zoltan Farkas commented on CAMEL-8994:
--------------------------------------
Hi Claus,
CL Hashmap 1.4.2 seems to fix the issue (remove the Padded types), and the code
is 1.6 compatible. I believe camel 2.14.3 uses something older, it would be
nice if we can upgrade.
I was able to take the latest ConcurrentLinkedHashap from Github which contains
the fix that removes the padded types that take up most of the memory, and use
it successfully with 2.14.3.
thank you
> Large memory use on Large core count(512) servers
> -------------------------------------------------
>
> Key: CAMEL-8994
> URL: https://issues.apache.org/jira/browse/CAMEL-8994
> Project: Camel
> Issue Type: Improvement
> Components: camel-core
> Affects Versions: 2.14.3
> Reporter: Zoltan Farkas
> Assignee: Claus Ibsen
> Fix For: 2.16.0, 2.15.3, 2.14.4
>
>
> ConcurrentLinkedHashmap get huge on these servers (200 +Mb)
> There are some CPU related defaults + Some padding that are the source of
> this, more discussion at:
> https://github.com/ben-manes/concurrentlinkedhashmap/issues/43
> upgrading to a newer version should remove the padding and reduce memory
> usage...
> Another thing I noticed that is related, is that LRUCache
> default initial size is equal with maximum size... which is 1000, would it
> make more sense to have a lower default for initialSize? (64)
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
