I guess the question is: does it make sense to have OL with RC? Seems like this combo
ignores concurrent changes, so maybe the isolation level default should change depending
on whether OL or PL is used?
On Sep 27, 2012, at 6:09 PM, Mircea Markus <mircea.markus(a)jboss.com> wrote:
As we name the transactions optimistic, wouldn't' it make
more sense to have WSK functionality enabled by default? (even though that would imply
using repeatable_read).
I mean as a user I kind of expect the tx to fail for optimistic transactions when
there's a write skew...
On 27 Sep 2012, at 16:12, Manik Surtani wrote:
> WSK only works for repeatable read. By default we use read committed.
>
> On 27 Sep 2012, at 16:07, Mircea Markus <mircea.markus(a)jboss.com> wrote:
>
>> Hi,
>>
>> By default our optimistic transactions don't have writeSkewCheck enabled so
their behaviour is counterintuitive for the user. That's because they don't fault
the commit in the case of a concurrent change.
>> Anyone can remember why these defaults are being used?
>> I'd rather have optimistic transactions with WSK enabled by default so that
we won't confuse users.
>>
>> Cheers,
>> --
>> Mircea Markus
>> Infinispan lead (
www.infinispan.org)
>>
>>
>>
>>
>> _______________________________________________
>> infinispan-dev mailing list
>> infinispan-dev(a)lists.jboss.org
>>
https://lists.jboss.org/mailman/listinfo/infinispan-dev
>
> --
> Manik Surtani
> manik(a)jboss.org
>
twitter.com/maniksurtani
>
> Platform Architect, JBoss Data Grid
>
http://red.ht/data-grid
>
> _______________________________________________
> infinispan-dev mailing list
> infinispan-dev(a)lists.jboss.org
>
https://lists.jboss.org/mailman/listinfo/infinispan-dev
Cheers,
--
Mircea Markus
Infinispan lead (
www.infinispan.org)
_______________________________________________
infinispan-dev mailing list
infinispan-dev(a)lists.jboss.org
https://lists.jboss.org/mailman/listinfo/infinispan-dev