I basically agree with Sanne here that having the prefix isolated opens
space for performance improvements, though if certain call is prefixed,
RegionFactory can always drop that prefix. The important thing is to
mention if the region name is prefixed or not in javadocs. I would also
prefer if *all* region names that RegionFactory is supposed to access
followed the same strategy (prefixed/not-prefixed).
5.1 included method
QueryResultsRegion buildQueryResultsRegion(String regionName,
Properties properties)
where StandardQueryCache did the prefix for us, the change in 5.3 to
QueryResultsRegion buildQueryResultsRegion(String regionName,
SessionFactoryImplementor sessionFactory)
does not indicate that the prefix won't be there and the javadoc is
missing completely.
Also, DomainDataRegionConfig.getRegionName() has no javadoc and
RegionFactory does not add the prefix.
@Steve could you please amend the javadoc and confirm if RF should
prefix region names?
@Sanne while cache manager per deployment is an obvious way to
distinguish regions@deployments, CM holds some heavy resources (e.g.
threads) - so I while we are supposed to scale number of caches up to
thousands (and we've fixed some problems that arise when you have for
example ~3k caches), ATM you're not supposed to scale CMs. And I don't
think that we'll focus in this direction since the bright future lies in
microservices :)
Radim
On 05/17/2018 12:23 PM, Sanne Grinovero wrote:
On 17 May 2018 at 11:11, Sanne Grinovero <sanne(a)hibernate.org>
wrote:
> I think this is the RegionFactory's responsibility, as Hibernate ORM
> alone can't know if this is necessary.
>
> Prefixing is one of many options to isolate caches; a Cache technology
> might wish to use a different approach by implementing a custom
> `org.hibernate.cache.spi.CacheKeysFactory`.
Hum, I regret how I wrote the above paragraph.
I actually meant that a Cache technology could implement isolation in
many different ways.
Using a custom `CacheKeysFactory` is just an example, there are many
other options as well. In fact, I believe a good choice for
application servers would be to have an independent CacheManager for
each deployment. Then it can safely inspect the deployment, see if
there are multiple Persistence Units using the same caching
technology, and implement further isolation only if necessary.
These thoughts are a consequence of a chat I had with Paul Ferraro
from the WildFly team, as he mentioned the size of the keys becoming
problematic with all the additional prefixing they need to apply. I
hope this will make it possible to e.g. create an "as small as
possible" unique identifier for the deployments to replace the
deployment name during serialization (to identify the CacheManager)
followed by a numeric id indexing the PUs within the deployment - or
nothing altogether in case of single PUs.
Of course, I don't expect that to be needed right away; the
RegionFactory could just do good old prefixing for now but I hope
we'll explore such improvements in the near future.
Thanks,
Sanne
> Not least the server / deployer might be able to hint the Cache
> provider to tell if isolation is at all necessary.
>
> In conclusion, by having Hibernate ORM not messing with prefixes
> allows other technologies to implement more efficient solutions. Our
> own code also ends up being more efficient by not needing to add a
> prefix during each and every access to the cache.
>
> Thanks,
> Sanne
>
> On 17 May 2018 at 06:34, Gail Badner <gbadner(a)redhat.com> wrote:
>> I see that cache region names are not being prefixed in 5.3.
>>
>> EnabledCaching sets the TimestampsRegion region name
>> to TimestampsRegion.class.getName(), and sets the QueryResultsRegion region
>> name to QueryResultsRegion.class.getName(). [1]
>>
>> Without a prefix, WildFly is failing intermittently when there are 2
>> persistence units with the query cache enabled due to:
>>
>> org.infinispan.commons.CacheConfigurationException: ISPN000453: Attempt to
>> define configuration for cache org.hibernate.cache.spi.TimestampsRegion
>> which already exists
>>
>> Entity region names are not being prefixed either.
>>
>> Should they be prefixed by Hibernate or by the RegionFactory?
>>
>> Regards,
>> Gail
>>
>> [1]
>>
https://github.com/hibernate/hibernate-orm/blob/master/hibernate-core/src...
>> _______________________________________________
>> hibernate-dev mailing list
>> hibernate-dev(a)lists.jboss.org
>>
https://lists.jboss.org/mailman/listinfo/hibernate-dev
--
Radim Vansa <rvansa(a)redhat.com>
JBoss Performance Team