Bruce Schuchardt created GEODE-5269:
---------------------------------------
Summary: CommitConflictException after TransactionInDoubtException
Key: GEODE-5269
URL: https://issues.apache.org/jira/browse/GEODE-5269
Project: Geode
Issue Type: Test
Components: client/server, transactions
Reporter: Bruce Schuchardt
If a client initiates a transaction commit and gets a
TransactionInDoubtException because the server hosting the transaction is
shutting down the client can easily get a CommitConflictException if it
attempts another transaction using one or more of the same keys as the in-doubt
transaction.
{noformat}
org.apache.geode.cache.CommitConflictException: Concurrent transaction commit
detected The key Object_130 in region /testRegion was being modified by
another transaction locally. org.apache.geode.cache.CommitConflictException:
Concurrent transaction commit detected The key Object_130 in region
/testRegion was being modified by another transaction locally.
at
org.apache.geode.internal.cache.locks.TXLockServiceImpl.txLock(TXLockServiceImpl.java:137)
at
org.apache.geode.internal.cache.TXLockRequest.obtain(TXLockRequest.java:88)
at
org.apache.geode.internal.cache.TXState.reserveAndCheck(TXState.java:335)
at org.apache.geode.internal.cache.TXState.commit(TXState.java:406)
at
org.apache.geode.internal.cache.TXStateProxyImpl.commit(TXStateProxyImpl.java:226)
at
org.apache.geode.internal.cache.TXManagerImpl.commit(TXManagerImpl.java:405)
at
org.apache.geode.internal.cache.tier.sockets.command.CommitCommand.cmdExecute(CommitCommand.java:82)
at
org.apache.geode.internal.cache.tier.sockets.BaseCommand.execute(BaseCommand.java:158)
at
org.apache.geode.internal.cache.tier.sockets.ServerConnection.doNormalMsg(ServerConnection.java:869)
at
org.apache.geode.internal.cache.tier.sockets.OriginalServerConnection.doOneMessage(OriginalServerConnection.java:77)
at
org.apache.geode.internal.cache.tier.sockets.ServerConnection.run(ServerConnection.java:1248)
at
java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1149)
at
java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:624)
at
org.apache.geode.internal.cache.tier.sockets.AcceptorImpl$4$1.run(AcceptorImpl.java:644)
at java.lang.Thread.run(Thread.java:748)
util.TestException: Unexpected org.apache.geode.cache.CommitConflictException:
Concurrent transaction commit detected The key Object_130 in region
/testRegion was being modified by another transaction locally.
org.apache.geode.cache.CommitConflictException: Concurrent transaction commit
detected The key Object_130 in region /testRegion was being modified by
another transaction locally.
at
org.apache.geode.internal.cache.locks.TXLockServiceImpl.txLock(TXLockServiceImpl.java:137)
at
org.apache.geode.internal.cache.TXLockRequest.obtain(TXLockRequest.java:88)
at
org.apache.geode.internal.cache.TXState.reserveAndCheck(TXState.java:335)
at org.apache.geode.internal.cache.TXState.commit(TXState.java:406)
at
org.apache.geode.internal.cache.TXStateProxyImpl.commit(TXStateProxyImpl.java:226)
at
org.apache.geode.internal.cache.TXManagerImpl.commit(TXManagerImpl.java:405)
at
org.apache.geode.internal.cache.tier.sockets.command.CommitCommand.cmdExecute(CommitCommand.java:82)
at
org.apache.geode.internal.cache.tier.sockets.BaseCommand.execute(BaseCommand.java:158)
at
org.apache.geode.internal.cache.tier.sockets.ServerConnection.doNormalMsg(ServerConnection.java:869)
at
org.apache.geode.internal.cache.tier.sockets.OriginalServerConnection.doOneMessage(OriginalServerConnection.java:77)
at
org.apache.geode.internal.cache.tier.sockets.ServerConnection.run(ServerConnection.java:1248)
at
java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1149)
at
java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:624)
at
org.apache.geode.internal.cache.tier.sockets.AcceptorImpl$4$1.run(AcceptorImpl.java:644)
at java.lang.Thread.run(Thread.java:748)
{noformat}
This is apparently due to the hosting server still being around and holding the
locks. If you wait for a bit and then try the transaction again it'll succeed
because the server has finished shutting down and has released the locks.
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
