merlimat opened a new pull request, #16660:
URL: https://github.com/apache/pulsar/pull/16660

   ### Motivation
   
   There is a sequence of events that can lead to a deadlock in the key-shared 
dispatcher.
   
   
   ```
   "BookKeeperClientWorker-OrderedExecutor-6-0":
        at 
org.apache.bookkeeper.mledger.impl.ManagedCursorImpl.readOperationCompleted(ManagedCursorImpl.java:2843)
        - waiting to lock <0x0000000684151218> (a java.util.ArrayDeque)
        at 
org.apache.bookkeeper.mledger.impl.OpReadEntry.checkReadCompletion(OpReadEntry.java:153)
        at 
org.apache.bookkeeper.mledger.impl.OpReadEntry.readEntriesComplete(OpReadEntry.java:87)
        at 
org.apache.bookkeeper.mledger.impl.EntryCacheImpl.asyncReadEntry0(EntryCacheImpl.java:285)
        at 
org.apache.bookkeeper.mledger.impl.EntryCacheImpl.asyncReadEntry(EntryCacheImpl.java:243)
        at 
org.apache.bookkeeper.mledger.impl.ManagedLedgerImpl.asyncReadEntry(ManagedLedgerImpl.java:1935)
        at 
org.apache.bookkeeper.mledger.impl.ManagedLedgerImpl.internalReadFromLedger(ManagedLedgerImpl.java:1907)
        at 
org.apache.bookkeeper.mledger.impl.ManagedLedgerImpl.asyncReadEntries(ManagedLedgerImpl.java:1707)
        at 
org.apache.bookkeeper.mledger.impl.ManagedCursorImpl.asyncReadEntries(ManagedCursorImpl.java:659)
        at 
org.apache.bookkeeper.mledger.impl.ManagedCursorImpl.asyncReadEntries(ManagedCursorImpl.java:642)
        at 
org.apache.bookkeeper.mledger.impl.ManagedCursorImpl.asyncReadEntriesOrWait(ManagedCursorImpl.java:793)
        at 
org.apache.pulsar.broker.service.persistent.PersistentDispatcherMultipleConsumers.readMoreEntries(PersistentDispatcherMultipleConsumers.java:274)
        - locked <0x00000006841e8bb8> (a 
org.apache.pulsar.broker.service.persistent.PersistentStickyKeyDispatcherMultipleConsumers)
        at 
org.apache.pulsar.broker.service.persistent.PersistentStickyKeyDispatcherMultipleConsumers.sendMessagesToConsumers(PersistentStickyKeyDispatcherMultipleConsumers.java:318)
        at 
org.apache.pulsar.broker.service.persistent.PersistentDispatcherMultipleConsumers.readEntriesComplete(PersistentDispatcherMultipleConsumers.java:483)
        - locked <0x00000006841e8bb8> (a 
org.apache.pulsar.broker.service.persistent.PersistentStickyKeyDispatcherMultipleConsumers)
        at 
org.apache.bookkeeper.mledger.impl.OpReadEntry.lambda$checkReadCompletion$2(OpReadEntry.java:157)
        at 
org.apache.bookkeeper.mledger.impl.OpReadEntry$$Lambda$1111/0x00000008408e9c40.run(Unknown
 Source)
        at org.apache.bookkeeper.mledger.util.SafeRun$1.safeRun(SafeRun.java:32)
        at 
org.apache.bookkeeper.common.util.SafeRunnable.run(SafeRunnable.java:36)
   
   "pulsar-io-4-7":
        at 
org.apache.pulsar.broker.service.persistent.PersistentStickyKeyDispatcherMultipleConsumers.markDeletePositionMoveForward(PersistentStickyKeyDispatcherMultipleConsumers.java:395)
        - waiting to lock <0x00000006841e8bb8> (a 
org.apache.pulsar.broker.service.persistent.PersistentStickyKeyDispatcherMultipleConsumers)
        at 
org.apache.pulsar.broker.service.persistent.PersistentSubscription.notifyTheMarkDeletePositionMoveForwardIfNeeded(PersistentSubscription.java:469)
        at 
org.apache.pulsar.broker.service.persistent.PersistentSubscription.access$100(PersistentSubscription.java:83)
        at 
org.apache.pulsar.broker.service.persistent.PersistentSubscription$2.deleteComplete(PersistentSubscription.java:456)
        at 
org.apache.bookkeeper.mledger.impl.ManagedCursorImpl$19.markDeleteComplete(ManagedCursorImpl.java:2123)
        at 
org.apache.bookkeeper.mledger.impl.ManagedCursorImpl$MarkDeleteEntry.triggerComplete(ManagedCursorImpl.java:234)
        at 
org.apache.bookkeeper.mledger.impl.ManagedCursorImpl.internalMarkDelete(ManagedCursorImpl.java:1830)
        at 
org.apache.bookkeeper.mledger.impl.ManagedCursorImpl.internalAsyncMarkDelete(ManagedCursorImpl.java:1811)
        - locked <0x0000000684151218> (a java.util.ArrayDeque)
        at 
org.apache.bookkeeper.mledger.impl.ManagedCursorImpl.asyncDelete(ManagedCursorImpl.java:2120)
        at 
org.apache.pulsar.broker.service.persistent.PersistentSubscription.acknowledgeMessage(PersistentSubscription.java:380)
        at 
org.apache.pulsar.broker.service.Consumer.individualAckNormal(Consumer.java:434)
        at 
org.apache.pulsar.broker.service.Consumer.messageAcked(Consumer.java:393)
        at 
org.apache.pulsar.broker.service.ServerCnx.handleAck(ServerCnx.java:1457)
        at 
org.apache.pulsar.common.protocol.PulsarDecoder.channelRead(PulsarDecoder.java:145)
        at 
io.netty.channel.AbstractChannelHandlerContext.invokeChannelRead(AbstractChannelHandlerContext.java:379)
        at 
io.netty.channel.AbstractChannelHandlerContext.invokeChannelRead(AbstractChannelHandlerContext.java:365)
   ```
   
   ### Modification
   
   We need to break the change and release the original mutex before processing 
the notification.
   
   


-- 
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.

To unsubscribe, e-mail: [email protected]

For queries about this service, please contact Infrastructure at:
[email protected]

Reply via email to