/*
* 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.zookeeper;
import java.util.concurrent.atomic.AtomicBoolean;
import org.apache.commons.logging.Log;
import org.apache.commons.logging.LogFactory;
import org.apache.hadoop.classification.InterfaceAudience;
import org.apache.hadoop.classification.InterfaceStability;
import org.apache.hadoop.hbase.Stoppable;
import org.apache.hadoop.hbase.util.Bytes;
import org.apache.zookeeper.KeeperException;
/**
* Handles coordination of a single "leader" instance among many possible
* candidates. The first {@link ZKLeaderManager} to successfully create
* the given znode becomes the leader, allowing the instance to continue
* with whatever processing must be protected. Other {@link ZKLeaderManager}
* instances will wait to be notified of changes to the leader znode.
* If the current master instance fails, the ephemeral leader znode will
* be removed, and all waiting instances will be notified, with the race
* to claim the leader znode beginning all over again.
* @deprecated Not used
*/
@Deprecated
@InterfaceAudience.Public
@InterfaceStability.Evolving
public class ZKLeaderManager extends ZooKeeperListener {
private static Log LOG = LogFactory.getLog(ZKLeaderManager.class);
private final AtomicBoolean leaderExists = new AtomicBoolean();
private String leaderZNode;
private byte[] nodeId;
private Stoppable candidate;
public ZKLeaderManager(ZooKeeperWatcher watcher, String leaderZNode,
byte[] identifier, Stoppable candidate) {
super(watcher);
this.leaderZNode = leaderZNode;
this.nodeId = identifier;
this.candidate = candidate;
}
public void start() {
try {
watcher.registerListener(this);
String parent = ZKUtil.getParent(leaderZNode);
if (ZKUtil.checkExists(watcher, parent) < 0) {
ZKUtil.createWithParents(watcher, parent);
}
} catch (KeeperException ke) {
watcher.abort("Unhandled zk exception when starting", ke);
candidate.stop("Unhandled zk exception starting up: "+ke.getMessage());
}
}
@Override
public void nodeCreated(String path) {
if (leaderZNode.equals(path) && !candidate.isStopped()) {
handleLeaderChange();
}
}
@Override
public void nodeDeleted(String path) {
if (leaderZNode.equals(path) && !candidate.isStopped()) {
handleLeaderChange();
}
}
private void handleLeaderChange() {
try {
synchronized(leaderExists) {
if (ZKUtil.watchAndCheckExists(watcher, leaderZNode)) {
LOG.info("Found new leader for znode: "+leaderZNode);
leaderExists.set(true);
} else {
LOG.info("Leader change, but no new leader found");
leaderExists.set(false);
leaderExists.notifyAll();
}
}
} catch (KeeperException ke) {
watcher.abort("ZooKeeper error checking for leader znode", ke);
candidate.stop("ZooKeeper error checking for leader: "+ke.getMessage());
}
}
/**
* Blocks until this instance has claimed the leader ZNode in ZooKeeper
*/
public void waitToBecomeLeader() {
while (!candidate.isStopped()) {
try {
if (ZKUtil.createEphemeralNodeAndWatch(watcher, leaderZNode, nodeId)) {
// claimed the leader znode
leaderExists.set(true);
if (LOG.isDebugEnabled()) {
LOG.debug("Claimed the leader znode as '"+
Bytes.toStringBinary(nodeId)+"'");
}
return;
}
// if claiming the node failed, there should be another existing node
byte[] currentId = ZKUtil.getDataAndWatch(watcher, leaderZNode);
if (currentId != null && Bytes.equals(currentId, nodeId)) {
// claimed with our ID, but we didn't grab it, possibly restarted?
LOG.info("Found existing leader with our ID ("+
Bytes.toStringBinary(nodeId)+"), removing");
ZKUtil.deleteNode(watcher, leaderZNode);
leaderExists.set(false);
} else {
LOG.info("Found existing leader with ID: "+Bytes.toStringBinary(nodeId));
leaderExists.set(true);
}
} catch (KeeperException ke) {
watcher.abort("Unexpected error from ZK, stopping candidate", ke);
candidate.stop("Unexpected error from ZK: "+ke.getMessage());
return;
}
// wait for next chance
synchronized(leaderExists) {
while (leaderExists.get() && !candidate.isStopped()) {
try {
leaderExists.wait();
} catch (InterruptedException ie) {
LOG.debug("Interrupted waiting on leader", ie);
}
}
}
}
}
/**
* Removes the leader znode, if it is currently claimed by this instance.
*/
public void stepDownAsLeader() {
try {
synchronized(leaderExists) {
if (!leaderExists.get()) {
return;
}
byte[] leaderId = ZKUtil.getData(watcher, leaderZNode);
if (leaderId != null && Bytes.equals(nodeId, leaderId)) {
LOG.info("Stepping down as leader");
ZKUtil.deleteNodeFailSilent(watcher, leaderZNode);
leaderExists.set(false);
} else {
LOG.info("Not current leader, no need to step down");
}
}
} catch (KeeperException ke) {
watcher.abort("Unhandled zookeeper exception removing leader node", ke);
candidate.stop("Unhandled zookeeper exception removing leader node: "
+ ke.getMessage());
}
}
public boolean hasLeader() {
return leaderExists.get();
}
}