org.apache.hadoop.hbase.procedure2.LockAndQueue Maven / Gradle / Ivy
/*
* Licensed to the Apache Software Foundation (ASF) under one
* or more contributor license agreements. See the NOTICE file
* distributed with this work for additional information
* regarding copyright ownership. The ASF licenses this file
* to you under the Apache License, Version 2.0 (the
* "License"); you may not use this file except in compliance
* with the License. You may obtain a copy of the License at
*
* http://www.apache.org/licenses/LICENSE-2.0
*
* Unless required by applicable law or agreed to in writing, software
* distributed under the License is distributed on an "AS IS" BASIS,
* WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
* See the License for the specific language governing permissions and
* limitations under the License.
*/
package org.apache.hadoop.hbase.procedure2;
import java.util.function.Function;
import java.util.function.Predicate;
import java.util.stream.Stream;
import org.apache.yetus.audience.InterfaceAudience;
/**
* Locking for mutual exclusion between procedures. Used only by procedure framework internally.
* {@link LockAndQueue} has two purposes:
*
* - Acquire/release exclusive/shared locks.
* - Maintains a list of procedures waiting on this lock. {@link LockAndQueue} extends
* {@link ProcedureDeque} class. Blocked Procedures are added to our super Deque. Using inheritance
* over composition to keep the Deque of waiting Procedures is unusual, but we do it this way
* because in certain cases, there will be millions of regions. This layout uses less memory.
*
*
* NOT thread-safe. Needs external concurrency control: e.g. uses in MasterProcedureScheduler are
* guarded by schedLock().
* There is no need of 'volatile' keyword for member variables because of memory synchronization
* guarantees of locks (see 'Memory Synchronization',
* http://docs.oracle.com/javase/8/docs/api/java/util/concurrent/locks/Lock.html)
* We do not implement Lock interface because we need exclusive and shared locking, and also because
* try-lock functions require procedure id.
* We do not use ReentrantReadWriteLock directly because of its high memory overhead.
*/
@InterfaceAudience.Private
public class LockAndQueue implements LockStatus {
private final Function> procedureRetriever;
private final ProcedureDeque queue = new ProcedureDeque();
private Procedure> exclusiveLockOwnerProcedure = null;
private int sharedLock = 0;
// ======================================================================
// Lock Status
// ======================================================================
public LockAndQueue(Function> procedureRetriever) {
this.procedureRetriever = procedureRetriever;
}
@Override
public boolean hasExclusiveLock() {
return this.exclusiveLockOwnerProcedure != null;
}
@Override
public boolean hasLockAccess(Procedure> proc) {
if (exclusiveLockOwnerProcedure == null) {
return false;
}
long lockOwnerId = exclusiveLockOwnerProcedure.getProcId();
if (proc.getProcId() == lockOwnerId) {
return true;
}
if (!proc.hasParent()) {
return false;
}
// fast path to check root procedure
if (proc.getRootProcId() == lockOwnerId) {
return true;
}
// check ancestors
for (Procedure> p = proc;;) {
if (p.getParentProcId() == lockOwnerId) {
return true;
}
p = procedureRetriever.apply(p.getParentProcId());
if (p == null || !p.hasParent()) {
return false;
}
}
}
@Override
public Procedure> getExclusiveLockOwnerProcedure() {
return exclusiveLockOwnerProcedure;
}
@Override
public int getSharedLockCount() {
return sharedLock;
}
// ======================================================================
// try/release Shared/Exclusive lock
// ======================================================================
/** Returns whether we have succesfully acquired the shared lock. */
public boolean trySharedLock(Procedure> proc) {
if (hasExclusiveLock() && !hasLockAccess(proc)) {
return false;
}
// If no one holds the xlock, then we are free to hold the sharedLock
// If the parent proc or we have already held the xlock, then we return true here as
// xlock is more powerful then shared lock.
sharedLock++;
return true;
}
/** Returns whether we should wake the procedures waiting on the lock here. */
public boolean releaseSharedLock() {
// hasExclusiveLock could be true, it usually means we acquire shared lock while we or our
// parent have held the xlock. And since there is still an exclusive lock, we do not need to
// wake any procedures.
return --sharedLock == 0 && !hasExclusiveLock();
}
public boolean tryExclusiveLock(Procedure> proc) {
if (isLocked()) {
return hasLockAccess(proc);
}
exclusiveLockOwnerProcedure = proc;
return true;
}
/** Returns whether we should wake the procedures waiting on the lock here. */
public boolean releaseExclusiveLock(Procedure> proc) {
if (
exclusiveLockOwnerProcedure == null
|| exclusiveLockOwnerProcedure.getProcId() != proc.getProcId()
) {
// We are not the lock owner, it is probably inherited from the parent procedures.
return false;
}
exclusiveLockOwnerProcedure = null;
// This maybe a bit strange so let me explain. We allow acquiring shared lock while the parent
// proc or we have already held the xlock, and also allow releasing the locks in any order, so
// it could happen that the xlock is released but there are still some procs holding the shared
// lock.
// In HBase, this could happen when a proc which holdLock is false and schedules sub procs which
// acquire the shared lock on the same lock. This is because we will schedule the sub proces
// before releasing the lock, so the sub procs could call acquire lock before we releasing the
// xlock.
return sharedLock == 0;
}
public boolean isWaitingQueueEmpty() {
return queue.isEmpty();
}
public Procedure> removeFirst() {
return queue.removeFirst();
}
public void addLast(Procedure> proc) {
queue.addLast(proc);
}
public int wakeWaitingProcedures(ProcedureScheduler scheduler) {
int count = queue.size();
// wakeProcedure adds to the front of queue, so we start from last in the waitQueue' queue, so
// that the procedure which was added first goes in the front for the scheduler queue.
scheduler.addFront(queue.descendingIterator());
queue.clear();
return count;
}
@SuppressWarnings("rawtypes")
public Stream filterWaitingQueue(Predicate predicate) {
return queue.stream().filter(predicate);
}
@Override
public String toString() {
return "exclusiveLockOwner=" + (hasExclusiveLock() ? getExclusiveLockProcIdOwner() : "NONE")
+ ", sharedLockCount=" + getSharedLockCount() + ", waitingProcCount=" + queue.size();
}
}