Terracotta Discussion Forums (LEGACY READ-ONLY ARCHIVE)

ljacomet wrote:

Versions of Ehcache above 2.6 no longer use an overflow model, but a tiering model.

In this new model, when you have a disk tier, it will always contain all cache entries.

Sorry, wrote that too fast: it needs to say versions of Ehcache 2.6 and above

Versions of Ehcache above 2.6 no longer use an overflow model, but a tiering model.

In this new model, when you have a disk tier, it will always contain all cache entries.

Hi,

Have a look at the following, it should help:
http://quartz-scheduler.org/documentation/quartz-2.x/tutorials/tutorial-lesson-06

Regards,

Can you make sure that there is not another Ehcache jar present in your classpath?

Also I would recommend using the latest 2.6.x version to benefit from bug fixes and improvements.

oxygenan wrote:

We have a requirement to write to the file only when the element is moved out of the cache. How can we do this?

Look into cache event listeners for this, as the file writing will have to happen outside Ehcache.

oxygenan wrote:

And is cache.put( new Element( key, value)); a good way of updating cache ?

Not sure I understand the question. For writing content to the cache, you have put and replace methods, see API doc.
You must not rely on mutating the value outside of the cache and hoping the cached value is updated as well. This would be looking for trouble.

My guess, without looking at code, would be that you are hitting a timeout in the replication layer.
With the static list, it will most likely try to reach each server, including the ones down, and so wait for the timeout to occur on any down server.
The fact that you regain this time when removing down servers supports this theory.

Hello,

You will want to use BigMemory Go for a single cache with off heap memory or BigMemory Max for a distributed cache.
You then need to request a trial license by following the links on respective pages.

We had a look and the issue is an Ehcache one.

We created EHC-1079 and will be working on a fix.

Thanks for the report.

Hi,

Which version of Ehcache are you using?
Did you try upgrading to a version where the fix for EHC-937 is available?

Could you please file an issue at https://jira.terracotta.org/jira/browse/EHC ?

Thanks

Terracotta server are started using shell scripts:
* start-tc-server.sh
* stop-tc-server.sh

Based on that, you will get better advice in forums for your linux distribution on how to create a service.

Yes you are right, I did not read the whole exception message closely enough.

You will need to look into class loaders to understand why Hibernate code would not be visible from Ehcache class loaders.

This is probably related to https://hibernate.atlassian.net/browse/HHH-8732

You will find a workaround here: http://forums.terracotta.org/forums/posts/list/8783.page

Some points worth checking:
- other version of ehcache(-ee) jar on the classpath? (but would need to be quite old)
- jrmc means JRockit I assume - can you try with Oracle JVM?
- was the jar tampered and the class is effectively no longer in it?

There is no longer an open-source version for Terracotta in the 4.x line. Check the website for the different offerings.