ISPN-1910 - State transfer should not force all invocations to be synchronous

[danberindei]

Also the t_1910_51 branch for 5.1.x.

https://issues.jboss.org/browse/ISPN-1910

I removed this limitation and replay support altogether, since it was no longer used.

[maniksurtani]

Will be good if Bela can run his benchmarks after this fix and see how
things perform. :)

Thanks for sorting this do quickly, Dan.

Sent from my mobile phone

On 13 Mar 2012, at 19:41, danberindei
reply@reply.github.com
wrote:

Also the t_1910_51 branch for 5.1.x.

https://issues.jboss.org/browse/ISPN-1910

I removed this limitation and replay support altogether, since it was no longer used.

You can merge this Pull Request by running:

git pull https://github.com/danberindei/infinispan t_1910_m

Or you can view, comment on it, or merge it online at:

#998

-- Commit Summary --

• ISPN-1910 - State transfer should not force all invocations to be synchronous

-- File Changes --

M core/src/main/java/org/infinispan/cacheviews/CacheViewsManagerImpl.java (13)
M core/src/main/java/org/infinispan/manager/DefaultCacheManager.java (2)
M core/src/main/java/org/infinispan/marshall/jboss/ExternalizerTable.java (4)
D core/src/main/java/org/infinispan/remoting/responses/ExtendedResponse.java (84)
D core/src/main/java/org/infinispan/remoting/responses/RequestIgnoredResponse.java (76)
M core/src/main/java/org/infinispan/remoting/rpc/RpcManagerImpl.java (2)
M core/src/main/java/org/infinispan/remoting/transport/Transport.java (4)
M core/src/main/java/org/infinispan/remoting/transport/jgroups/CommandAwareRpcDispatcher.java (95)
M core/src/main/java/org/infinispan/remoting/transport/jgroups/JGroupsTransport.java (13)
M core/src/test/java/org/infinispan/api/mvcc/PutForExternalReadTest.java (2)
M core/src/test/java/org/infinispan/invalidation/BaseInvalidationTest.java (2)
M core/src/test/java/org/infinispan/jmx/RpcManagerMBeanTest.java (4)
M core/src/test/java/org/infinispan/marshall/VersionAwareMarshallerTest.java (15)
M core/src/test/java/org/infinispan/replication/SyncReplTest.java (4)
M spring/src/test/java/org/infinispan/spring/mock/MockTransport.java (3)

-- Patch Links --

https://github.com/infinispan/infinispan/pull/998.patch
https://github.com/infinispan/infinispan/pull/998.diff

---

Reply to this email directly or view it on GitHub:
#998

[tristantarrant]

Dan, on 5.1.x I'm getting this: Entry under key [test146] was [null] but expected [146]" type="java.lang.AssertionError">java.lang.AssertionError: Entry under key [test146] was [null] but expected [146]
at org.infinispan.statetransfer.StateTransferFunctionalTest.thirdWritingCacheTest(StateTransferFunctionalTest.java:338)
at org.infinispan.statetransfer.StateTransferFunctionalTest.testSTWithThirdWritingNonTxCache(StateTransferFunctionalTest.java:251)

[tristantarrant]

And on master I get the following failures:
Failed tests:
testInitiatorNodeCrashesBeforeCommit(org.infinispan.lock.singlelock.replicated.pessimistic.InitiatorCrashPessimisticReplTest)
testStateTransfer(org.infinispan.container.versioning.VersionedDistStateTransferTest)

[danberindei]

@tristantarrant, I've reproduced the StateTransferFunctionalTest failure both with and without this commit.
It looks like I had assumed in StateTransferLockInterceptor that I would catch a SuspectException when a target node left the cluster, but instead I'm catching a CacheException. I've created https://issues.jboss.org/browse/ISPN-1931 for this issue, in the meantime I think it's safe to integrate this pull req.