I wrote the original test, to simulate a transactional write occuring (still in prepare
phase) while a node joins. So to this end, I think we can safely run the test without L1
caching as the behaviour we are trying to test is still correct.
However I'm puzzled that the L1 invalidation kills this - the L1 invalidation should
be a fail-fast mechanism.
On 19 Nov 2010, at 17:20, Vladimir Blagojevic wrote:
Hi,
SingleJoinTest transactional variant currently fails. But it is destined
to fail as is! Let me explain. We have four nodes {A,B,C,D} and another
node E joins. Just before E joins a tx T is started on one of the
existing nodes with a write to key K. There is essentially a latch in
the prepare of T that is opened only after E succefully joins upon join
rehash. As part of join rehash E needs to invalidate invalid L1 keys and
during that process it tries to lock ..you guessed it key K. Since T
holds a lock on K already eventually E join rehash task times out and as
a result we get a failed test.
I did not write this test and I would not like to change the original
writer's test case intent. What are our options?
Cheers,
Vladimir
_______________________________________________
infinispan-dev mailing list
infinispan-dev(a)lists.jboss.org
https://lists.jboss.org/mailman/listinfo/infinispan-dev
--
Manik Surtani
manik(a)jboss.org
Lead, Infinispan
Lead, JBoss Cache
http://www.infinispan.org
http://www.jbosscache.org