configure : instance-pool/instance-cache

[cleconte]

Hi,

I am trying to configure my stateful session bean. But I don't understand the difference between "the instance-pool" and the "instance cache". When I change the parameters (cache-policy-conf) as "min-capacity", "max-capacity" and "max-bean-age" I see that the ejbActivate method and ejbPassivate are called. But when I change the parameter (container-pool-conf) as "MaximumZise" I do not see change.

<cache-policy-conf>
<min-capacity>1</min-capacity> <max-capacity>3</max-capacity>
<remover-period>1800</remover-period>
<max-bean-life>1800</max-bean-life>
<overager-period>300</overager-period>
<max-bean-age>600</max-bean-age>
<resizer-period>400</resizer-period>
<max-cache-miss-period>60</max-cache-miss-period>
<min-cache-miss-period>1</min-cache-miss-period>
<cache-load-factor>0.75</cache-load-factor>
</cache-policy-conf>
</container-cache-conf>
<container-pool-conf>
1 1 true
</container-pool-conf>


Can you give an explanation between instance-pool and instance-cache.

Thanks.

Christophe.

[adrian.brock]

The pool is a collection of unassociated beans.
When a request comes in, a bean is selected from the
pool and attached to a bean context.
When the request finishes, the bean is removed from the
context and placed back in the pool, unless...

If you have a cache, the bean may be placed in the cache
and not returned to the pool after the request. In future
when somebody asks for the same key (either the internal
session key for Stateful Session beans or the primary key
for entity beans) the same bean is reused.

If the cache is too full or an error occurs while using
the bean, the bean can be passivated which involves
removing the context and maybe returning the bean
to the pool.

In simpler terms:
The pool is a collection of "raw" EJBs with no data
(all equivalent)
The cache is a collection of EJBs associated with
some data (none are equivalent to each other).

Regards,
Adrian

[demiante]

Hello Adrian and folks,
I read the Jboss Administration (payed documentation) and looked in the net but I didn't get this questions answered entirely yeat:

1)What is cache miss?
2)When it happens? in what cases?

My only hint is when the cache is full, like number-of-beans-activate = max-capacity-of-cache and another bean must to be create so the LRU bean must to be passivate.
If the cache miss only occurs in this case that i mentioned so why the resizer task limit the enlargement of the cache to the size of the <max-capacity> tag value?

I don't know if I get my questions understood but i appreciate any help,

Regards,
Vinicius