cocoon-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Ard Schrijvers (JIRA)" <j...@apache.org>
Subject [jira] Commented: (COCOON-1868) When using EHDefaultStore in overflow-to-disk=false mode, shutting the jvm down results in a spoolToDisk error
Date Thu, 22 Jun 2006 07:57:31 GMT
    [ http://issues.apache.org/jira/browse/COCOON-1868?page=comments#action_12417268 ] 

Ard Schrijvers commented on COCOON-1868:
----------------------------------------

You can only reproduce the bug, when there is actually at least one thing in the EHDefaultStore,
so starting cocoon and immediately shutting it down does not reproduce the bug. 

What I though just tested, is trying it with ehcache 1.2. Indeed, it seems to added to ehcache
1.2 that when overflow-to-disk is set to false, it does not try to spoolToDisk on a shut down
when diskPersistent is true. With ehcache 1.1, this was probably not tested, resulting in
the spoolToDisk error. 

So, the bug was fixed with ehcache 1.2. Though adding the diskpersistent parameter still adds
functionality: If you want the number of cachekeys  to be able to grow beyond the number defined
in maxobjects, you define "overflow-to-disk"=true. Then by default, when "diskpersitent" is
not defined, on a JVM shutdown, the part of the cache in memory is spooled to disk. 
Though, you might want to have a not diskPersistent cache (for example during development
you perhaps dont want the cache stored on disk when JVM shuts down and the cache gets populated
on a restart). With the patch, you are able (not obliged) to define how you want your ehcache
to behave.

I must though admit that with ehcache 1.2, since the bug is solved, only some marginal functionality
is added by this patch :-) 

COCOON-1694 was another fix, that is uncorrelated to the part I did. It was about the StoreJanitor
disposing all its caches on a cocoon shutdown. It did not check wether the cache was still
alive (it could have been stopped by another process). So, before cacheManager.shutdown()
is called, a check wether the cache is still alive is done. 

> When using EHDefaultStore in overflow-to-disk=false mode, shutting the jvm down results
in a spoolToDisk error
> --------------------------------------------------------------------------------------------------------------
>
>          Key: COCOON-1868
>          URL: http://issues.apache.org/jira/browse/COCOON-1868
>      Project: Cocoon
>         Type: Bug

>   Components: * Cocoon Core
>     Versions: 2.1.8
>     Reporter: Ard Schrijvers
>     Assignee: Antonio Gallardo
>     Priority: Minor
>  Attachments: EHDefaultStore.patch
>
> When using EHDefaultStore in overflow-to-disk=false mode, shutting the jvm down results
in a spoolToDisk error. This is because diskPersistent is not Parameterizable and always TRUE.
However, when overflow-to-disk is false (caching in memory only), when shutting the JVM down,
ehcache is trying to flush the cache to disk because diskPersistent is true. 
> Added diskPersistent parameter, which is FALSE when overflow-to-disk is false, and default
TRUE when overflow-to-disk is TRUE.

-- 
This message is automatically generated by JIRA.
-
If you think it was sent incorrectly contact one of the administrators:
   http://issues.apache.org/jira/secure/Administrators.jspa
-
For more information on JIRA, see:
   http://www.atlassian.com/software/jira


Mime
View raw message