[infinispan-dev] Improving batching docu WAS Re: transaction over hotrod

[Galder Zamarreño]

Mircea, thanks for updating http://community.jboss.org/docs/DOC-14901 but it's still confusing me and not all questions posted seem to be answered.

First of all, it seems to me that the info you added contradicts itself:

On one side, you say that batching should only be used when the only transaction participant is Infinispan cache. However, the latest paragraph talks about batching working fine with other resources via a cache loader, i.e. JTA datasource.

Also, can batching be used safely in a clustered environment where multiple Infinispan cache instances are involved? 

My questions on what happens with clustered caches remain unanswered in that wiki...

On Jul 22, 2010, at 3:14 AM, Mircea Markus wrote:

> 
> On 21 Jul 2010, at 14:49, Galder Zamarreño wrote:
> 
>> Aside from these questions, I think http://community.jboss.org/docs/DOC-14901 could do with having more information. Brian's http://community.jboss.org/wiki/BatchModeTransactionManager contains some info on what's the rationale behind batching and I suspect suspect some of that is useful in Infinispan as well.
>> 
>> On top of that, I think http://community.jboss.org/docs/DOC-14901 should answer:
>> - Are locks maintained in a batch?
>> - What happens at commit time in a local cache?
>> - What happens at commit time in a clustered cache? 1PC or 2PC
>> - What happens if a cache loader is involved? Are updates atomic on the cacheloader?
>> - When/Why would someone use batching instead of JTA transactions? And viceversa?
>> ...etc
> 
> https://jira.jboss.org/browse/ISPN-549
> 
> _______________________________________________
> infinispan-dev mailing list
> infinispan-dev at lists.jboss.org
> https://lists.jboss.org/mailman/listinfo/infinispan-dev

--
Galder Zamarreño
Sr. Software Engineer
Infinispan, JBoss Cache