GitHub user sk0x50 opened a pull request: https://github.com/apache/ignite/pull/2981
Ignite 6496: Client node does not reconnect to server node when the latter is restarted. The following scenario may result in deadlock on the client node: - start server node and one client - client node invokes IgniteQueue#take() method, that requires acquiring both GridCacheGateway#readLock and GridCacheQueueAdapter#readSem. - client node disconnects from the cluster for some reason (for example, server node was stopped) in that case, GridCacheQueueAdapter does not release readSem and, therefore, GridCacheGateway#readLock is also not released. - 'tcp-client-disco-msg-worker' hangs due to unable acquiring GridCacheGateway.writeLock ``` "tcp-client-disco-msg-worker-#10%datastructures.GridCacheQueueClientReconnect1%" #101 prio=5 os_prio=0 tid=0x0000000021634000 nid=0x49a8 waiting on condition [0x000000003a61e000] java.lang.Thread.State: WAITING (parking) at sun.misc.Unsafe.park(Native Method) parking to wait for <0x000000076fa988f8> (a java.util.concurrent.locks.ReentrantReadWriteLock$NonfairSync) at java.util.concurrent.locks.LockSupport.park(LockSupport.java:175) at java.util.concurrent.locks.AbstractQueuedSynchronizer.parkAndCheckInterrupt(AbstractQueuedSynchronizer.java:836) at java.util.concurrent.locks.AbstractQueuedSynchronizer.acquireQueued(AbstractQueuedSynchronizer.java:870) at java.util.concurrent.locks.AbstractQueuedSynchronizer.acquire(AbstractQueuedSynchronizer.java:1199) at java.util.concurrent.locks.ReentrantReadWriteLock$WriteLock.lock(ReentrantReadWriteLock.java:943) at org.apache.ignite.internal.util.StripedCompositeReadWriteLock$WriteLock.lock0(StripedCompositeReadWriteLock.java:154) at org.apache.ignite.internal.util.StripedCompositeReadWriteLock$WriteLock.lock(StripedCompositeReadWriteLock.java:123) at org.apache.ignite.internal.processors.cache.GridCacheGateway.writeLock(GridCacheGateway.java:278) at org.apache.ignite.internal.IgniteKernal.onDisconnected(IgniteKernal.java:3873) at org.apache.ignite.internal.managers.discovery.GridDiscoveryManager$4.onDiscovery0(GridDiscoveryManager.java:768) at org.apache.ignite.internal.managers.discovery.GridDiscoveryManager$4.onDiscovery(GridDiscoveryManager.java:573) locked <0x000000076f9f4048> (a java.lang.Object) at org.apache.ignite.spi.discovery.tcp.ClientImpl$MessageWorker.notifyDiscovery(ClientImpl.java:2415) at org.apache.ignite.spi.discovery.tcp.ClientImpl$MessageWorker.notifyDiscovery(ClientImpl.java:2394) at org.apache.ignite.spi.discovery.tcp.ClientImpl$MessageWorker.body(ClientImpl.java:1710) at org.apache.ignite.spi.IgniteSpiThread.run(IgniteSpiThread.java:62) ``` You can merge this pull request into a Git repository by running: $ git pull https://github.com/sk0x50/ignite ignite-6496 Alternatively you can review and apply these changes as the patch at: https://github.com/apache/ignite/pull/2981.patch To close this pull request, make a commit to your master/trunk branch with (at least) the following in the commit message: This closes #2981 ---- commit db7f1e497fe2d2502e910ab06e953e604768ed9d Author: Slava Koptilin <slava.kopti...@gmail.com> Date: 2017-10-31T15:29:09Z ignite-6496: release all IgniteQueue's semaphores in case of the client node disconnected commit e3eb26efd64c62d1de53a09f74b3f77491d2dd3d Author: Slava Koptilin <slava.kopti...@gmail.com> Date: 2017-11-03T14:01:25Z ignite-6496: added test ---- ---