Fixed deadlock on txn semaphore permit exhaustion #14131
Merged
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
merlimat
requested review from
Jason918,
rdhabalia,
jerrypeng,
hangc0276,
eolivelli,
codelipenghui,
BewareMyPower and
congbobo184
merlimat
added
doc-not-needed
eolivelli
approved these changes
Feb 5, 2022
hezhangjian
approved these changes
Feb 5, 2022
I remember the semaphore is for limiting the transaction buffer calls while reply transaction logs. @congbobo184 Please help double confirm. |
codelipenghui
approved these changes
Feb 7, 2022
codelipenghui
pushed a commit
that referenced
this pull request
Feb 10, 2022
### Motivation Removing semaphore on the end of transactions operations. The semaphore is not very useful here as we are already closing the transactions (backpressure should eventually be applied at the starting of the transactions). The semaphore here is being acquired from a BK callback thread and it causes a deadlock in broker when the semaphore is full, because the response that will release the permits on the semaphore are coming from either the same thread or a thread in the same condition. ``` sun.misc.Unsafe.park(Unsafe.java) java.util.concurrent.locks.LockSupport.park(LockSupport.java:175) java.util.concurrent.locks.AbstractQueuedSynchronizer.parkAndCheckInterrupt(AbstractQueuedSynchronizer.java:836) java.util.concurrent.locks.AbstractQueuedSynchronizer.doAcquireSharedInterruptibly(AbstractQueuedSynchronizer.java:997) java.util.concurrent.locks.AbstractQueuedSynchronizer.acquireSharedInterruptibly(AbstractQueuedSynchronizer.java:1304) java.util.concurrent.Semaphore.acquire(Semaphore.java:312) org.apache.pulsar.broker.transaction.buffer.impl.TransactionBufferHandlerImpl.canSendRequest(TransactionBufferHandlerImpl.java:216) org.apache.pulsar.broker.transaction.buffer.impl.TransactionBufferHandlerImpl.endTxnOnTopic(TransactionBufferHandlerImpl.java:93) org.apache.pulsar.broker.transaction.buffer.impl.TransactionBufferClientImpl.commitTxnOnTopic(TransactionBufferClientImpl.java:50) org.apache.pulsar.broker.TransactionMetadataStoreService.lambda$null$23(TransactionMetadataStoreService.java:484) org.apache.pulsar.broker.TransactionMetadataStoreService$$Lambda$1253.accept() java.util.ArrayList.forEach(ArrayList.java:1257) org.apache.pulsar.broker.TransactionMetadataStoreService.lambda$endTxnInTransactionBuffer$25(TransactionMetadataStoreService.java:481) org.apache.pulsar.broker.TransactionMetadataStoreService$$Lambda$1251.accept() java.util.concurrent.CompletableFuture.uniWhenComplete(CompletableFuture.java:760) java.util.concurrent.CompletableFuture.uniWhenCompleteStage(CompletableFuture.java:778) java.util.concurrent.CompletableFuture.whenComplete(CompletableFuture.java:2140) org.apache.pulsar.broker.TransactionMetadataStoreService.endTxnInTransactionBuffer(TransactionMetadataStoreService.java:458) org.apache.pulsar.broker.TransactionMetadataStoreService.lambda$null$11(TransactionMetadataStoreService.java:349) org.apache.pulsar.broker.TransactionMetadataStoreService$$Lambda$1309.accept() java.util.concurrent.CompletableFuture.uniAccept(CompletableFuture.java:656) java.util.concurrent.CompletableFuture$UniAccept.tryFire(CompletableFuture.java:632) java.util.concurrent.CompletableFuture.postComplete(CompletableFuture.java:474) java.util.concurrent.CompletableFuture.complete(CompletableFuture.java:1962) org.apache.pulsar.transaction.coordinator.impl.MLTransactionLogImpl$3.addComplete(MLTransactionLogImpl.java:160) org.apache.bookkeeper.mledger.impl.OpAddEntry.safeRun(OpAddEntry.java:228) org.apache.bookkeeper.common.util.SafeRunnable.run(SafeRunnable.java:36) java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1149) java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:624) io.netty.util.concurrent.FastThreadLocalRunnable.run(FastThreadLocalRunnable.java:30) java.lang.Thread.run(Thread.java:748) ``` (cherry picked from commit bea5bb8)
Nicklee007
pushed a commit
to Nicklee007/pulsar
that referenced
this pull request
Apr 20, 2022
### Motivation Removing semaphore on the end of transactions operations. The semaphore is not very useful here as we are already closing the transactions (backpressure should eventually be applied at the starting of the transactions). The semaphore here is being acquired from a BK callback thread and it causes a deadlock in broker when the semaphore is full, because the response that will release the permits on the semaphore are coming from either the same thread or a thread in the same condition. ``` sun.misc.Unsafe.park(Unsafe.java) java.util.concurrent.locks.LockSupport.park(LockSupport.java:175) java.util.concurrent.locks.AbstractQueuedSynchronizer.parkAndCheckInterrupt(AbstractQueuedSynchronizer.java:836) java.util.concurrent.locks.AbstractQueuedSynchronizer.doAcquireSharedInterruptibly(AbstractQueuedSynchronizer.java:997) java.util.concurrent.locks.AbstractQueuedSynchronizer.acquireSharedInterruptibly(AbstractQueuedSynchronizer.java:1304) java.util.concurrent.Semaphore.acquire(Semaphore.java:312) org.apache.pulsar.broker.transaction.buffer.impl.TransactionBufferHandlerImpl.canSendRequest(TransactionBufferHandlerImpl.java:216) org.apache.pulsar.broker.transaction.buffer.impl.TransactionBufferHandlerImpl.endTxnOnTopic(TransactionBufferHandlerImpl.java:93) org.apache.pulsar.broker.transaction.buffer.impl.TransactionBufferClientImpl.commitTxnOnTopic(TransactionBufferClientImpl.java:50) org.apache.pulsar.broker.TransactionMetadataStoreService.lambda$null$23(TransactionMetadataStoreService.java:484) org.apache.pulsar.broker.TransactionMetadataStoreService$$Lambda$1253.accept() java.util.ArrayList.forEach(ArrayList.java:1257) org.apache.pulsar.broker.TransactionMetadataStoreService.lambda$endTxnInTransactionBuffer$25(TransactionMetadataStoreService.java:481) org.apache.pulsar.broker.TransactionMetadataStoreService$$Lambda$1251.accept() java.util.concurrent.CompletableFuture.uniWhenComplete(CompletableFuture.java:760) java.util.concurrent.CompletableFuture.uniWhenCompleteStage(CompletableFuture.java:778) java.util.concurrent.CompletableFuture.whenComplete(CompletableFuture.java:2140) org.apache.pulsar.broker.TransactionMetadataStoreService.endTxnInTransactionBuffer(TransactionMetadataStoreService.java:458) org.apache.pulsar.broker.TransactionMetadataStoreService.lambda$null$11(TransactionMetadataStoreService.java:349) org.apache.pulsar.broker.TransactionMetadataStoreService$$Lambda$1309.accept() java.util.concurrent.CompletableFuture.uniAccept(CompletableFuture.java:656) java.util.concurrent.CompletableFuture$UniAccept.tryFire(CompletableFuture.java:632) java.util.concurrent.CompletableFuture.postComplete(CompletableFuture.java:474) java.util.concurrent.CompletableFuture.complete(CompletableFuture.java:1962) org.apache.pulsar.transaction.coordinator.impl.MLTransactionLogImpl$3.addComplete(MLTransactionLogImpl.java:160) org.apache.bookkeeper.mledger.impl.OpAddEntry.safeRun(OpAddEntry.java:228) org.apache.bookkeeper.common.util.SafeRunnable.run(SafeRunnable.java:36) java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1149) java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:624) io.netty.util.concurrent.FastThreadLocalRunnable.run(FastThreadLocalRunnable.java:30) java.lang.Thread.run(Thread.java:748) ```
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Motivation
Removing semaphore on the end of transactions operations. The semaphore is not very useful here as we are already closing the transactions (backpressure should eventually be applied at the starting of the transactions).
The semaphore here is being acquired from a BK callback thread and it causes a deadlock in broker when the semaphore is full, because the response that will release the permits on the semaphore are coming from either the same thread or a thread in the same condition.