[infinispan-dev] evict

[Mircea Markus]

Manik Surtani wrote:
> The evict call should not happen within the context of a transaction 
> (I don't see why this should ever happen).  It would ideally be called 
> by a maintenance thread (either our own eviction thread or a user 
> controlled one) but the sole purpose of this operation is to trim down 
> memory usage and not to actually change the state in the system (this 
> is why it doesn't affect cache stores).  If the javadocs around this 
> aren't as clear, could you please update them?
>
> Thanks
> Manik
I've updated the javadoc to say that the method should not be called 
from a tx. Also the TxInterceptor ignores this call now.
>
> On 30 Apr 2009, at 22:59, Mircea Markus wrote:
>
>> Hi,
>>
>> There is no comment in code next to "Cache.evict" method regarding 
>> its behavior with respect to transactions and replications.
>> How should it behave?
>> While i don't think it should be replicable (we do not want to 
>> replicate the evict process), how do we handle it in the context of 
>> Tx? Should we replicate it and handle it as a remove?
>>
>> Cheers,
>> Mircea
>> _______________________________________________
>> infinispan-dev mailing list
>> infinispan-dev at lists.jboss.org
>> https://lists.jboss.org/mailman/listinfo/infinispan-dev
>
> -- 
> Manik Surtani
> manik at jboss.org
> Lead, Infinispan
> Lead, JBoss Cache
> http://www.infinispan.org
> http://www.jbosscache.org
>
>
>
>