/**
* Ensure that the given NameNode marks the specified DataNode as entirely dead/expired.
*
* @param nn the NameNode to manipulate
* @param dnName the name of the DataNode
*/
public static void noticeDeadDatanode(NameNode nn, String dnName) {
FSNamesystem namesystem = nn.getNamesystem();
namesystem.writeLock();
try {
DatanodeManager dnm = namesystem.getBlockManager().getDatanodeManager();
HeartbeatManager hbm = dnm.getHeartbeatManager();
DatanodeDescriptor[] dnds = hbm.getDatanodes();
DatanodeDescriptor theDND = null;
for (DatanodeDescriptor dnd : dnds) {
if (dnd.getXferAddr().equals(dnName)) {
theDND = dnd;
}
}
Assert.assertNotNull("Could not find DN with name: " + dnName, theDND);
synchronized (hbm) {
DFSTestUtil.setDatanodeDead(theDND);
hbm.heartbeatCheck();
}
} finally {
namesystem.writeUnlock();
}
}
/**
* Stop decommissioning the specified datanode.
*
* @param node
*/
@VisibleForTesting
public void stopDecommission(DatanodeDescriptor node) {
if (node.isDecommissionInProgress() || node.isDecommissioned()) {
// Update DN stats maintained by HeartbeatManager
hbManager.stopDecommission(node);
// Over-replicated blocks will be detected and processed when
// the dead node comes back and send in its full block report.
if (node.isAlive) {
blockManager.processOverReplicatedBlocksOnReCommission(node);
}
// Remove from tracking in DecommissionManager
pendingNodes.remove(node);
decomNodeBlocks.remove(node);
} else {
LOG.trace("stopDecommission: Node {} in {}, nothing to do." + node, node.getAdminState());
}
}
/**
* Start decommissioning the specified datanode.
*
* @param node
*/
@VisibleForTesting
public void startDecommission(DatanodeDescriptor node) {
if (!node.isDecommissionInProgress() && !node.isDecommissioned()) {
// Update DN stats maintained by HeartbeatManager
hbManager.startDecommission(node);
// hbManager.startDecommission will set dead node to decommissioned.
if (node.isDecommissionInProgress()) {
for (DatanodeStorageInfo storage : node.getStorageInfos()) {
LOG.info(
"Starting decommission of {} {} with {} blocks", node, storage, storage.numBlocks());
}
node.decommissioningStatus.setStartTime(monotonicNow());
pendingNodes.add(node);
}
} else {
LOG.trace("startDecommission: Node {} in {}, nothing to do." + node, node.getAdminState());
}
}
@Test
public void testHeartbeatStopWatch() throws Exception {
Namesystem ns = Mockito.mock(Namesystem.class);
BlockManager bm = Mockito.mock(BlockManager.class);
Configuration conf = new Configuration();
long recheck = 2000;
conf.setLong(DFSConfigKeys.DFS_NAMENODE_HEARTBEAT_RECHECK_INTERVAL_KEY, recheck);
HeartbeatManager monitor = new HeartbeatManager(ns, bm, conf);
monitor.restartHeartbeatStopWatch();
assertFalse(monitor.shouldAbortHeartbeatCheck(0));
// sleep shorter than recheck and verify shouldn't abort
Thread.sleep(100);
assertFalse(monitor.shouldAbortHeartbeatCheck(0));
// sleep longer than recheck and verify should abort unless ignore delay
Thread.sleep(recheck);
assertTrue(monitor.shouldAbortHeartbeatCheck(0));
assertFalse(monitor.shouldAbortHeartbeatCheck(-recheck * 3));
// ensure it resets properly
monitor.restartHeartbeatStopWatch();
assertFalse(monitor.shouldAbortHeartbeatCheck(0));
}
public void run() {
while (Server.leaderNode == HeartbeatManager.nodeId) {
if (queue.isEmpty()) {
continue;
} else {
String node = queue.poll();
if (node == HeartbeatManager.nodeId) HeartbeatListener.leadack = 1;
else HeartbeatManager.broadcast(node, "1");
System.out.println("Lead ack sent");
while (HeartbeatListener.leadack != 2) {
continue;
}
System.out.println("Lead complete");
}
}
}