use of org.apache.solr.common.cloud.ZkNodeProps in project lucene-solr by apache.
the class OverseerCollectionMessageHandler method balanceProperty.
private void balanceProperty(ClusterState clusterState, ZkNodeProps message, NamedList results) throws KeeperException, InterruptedException {
if (StringUtils.isBlank(message.getStr(COLLECTION_PROP)) || StringUtils.isBlank(message.getStr(PROPERTY_PROP))) {
throw new SolrException(ErrorCode.BAD_REQUEST, "The '" + COLLECTION_PROP + "' and '" + PROPERTY_PROP + "' parameters are required for the BALANCESHARDUNIQUE operation, no action taken");
}
SolrZkClient zkClient = zkStateReader.getZkClient();
DistributedQueue inQueue = Overseer.getStateUpdateQueue(zkClient);
Map<String, Object> propMap = new HashMap<>();
propMap.put(Overseer.QUEUE_OPERATION, BALANCESHARDUNIQUE.toLower());
propMap.putAll(message.getProperties());
inQueue.offer(Utils.toJSON(new ZkNodeProps(propMap)));
}
use of org.apache.solr.common.cloud.ZkNodeProps in project lucene-solr by apache.
the class OverseerTaskProcessor method run.
@Override
public void run() {
log.debug("Process current queue of overseer operations");
LeaderStatus isLeader = amILeader();
while (isLeader == LeaderStatus.DONT_KNOW) {
log.debug("am_i_leader unclear {}", isLeader);
// not a no, not a yes, try ask again
isLeader = amILeader();
}
String oldestItemInWorkQueue = null;
// hasLeftOverItems - used for avoiding re-execution of async tasks that were processed by a previous Overseer.
// This variable is set in case there's any task found on the workQueue when the OCP starts up and
// the id for the queue tail is used as a marker to check for the task in completed/failed map in zk.
// Beyond the marker, all tasks can safely be assumed to have never been executed.
boolean hasLeftOverItems = true;
try {
oldestItemInWorkQueue = workQueue.getTailId();
} catch (KeeperException e) {
// We don't need to handle this. This is just a fail-safe which comes in handy in skipping already processed
// async calls.
SolrException.log(log, "", e);
} catch (InterruptedException e) {
Thread.currentThread().interrupt();
}
if (oldestItemInWorkQueue == null)
hasLeftOverItems = false;
else
log.debug("Found already existing elements in the work-queue. Last element: {}", oldestItemInWorkQueue);
try {
prioritizer.prioritizeOverseerNodes(myId);
} catch (Exception e) {
if (!zkStateReader.getZkClient().isClosed()) {
log.error("Unable to prioritize overseer ", e);
}
}
// TODO: Make maxThreads configurable.
this.tpe = new ExecutorUtil.MDCAwareThreadPoolExecutor(5, MAX_PARALLEL_TASKS, 0L, TimeUnit.MILLISECONDS, new SynchronousQueue<Runnable>(), new DefaultSolrThreadFactory("OverseerThreadFactory"));
try {
while (!this.isClosed) {
try {
isLeader = amILeader();
if (LeaderStatus.NO == isLeader) {
break;
} else if (LeaderStatus.YES != isLeader) {
log.debug("am_i_leader unclear {}", isLeader);
// not a no, not a yes, try asking again
continue;
}
log.debug("Cleaning up work-queue. #Running tasks: {}", runningTasks.size());
cleanUpWorkQueue();
printTrackingMaps();
boolean waited = false;
while (runningTasks.size() > MAX_PARALLEL_TASKS) {
synchronized (waitLock) {
//wait for 100 ms or till a task is complete
waitLock.wait(100);
}
waited = true;
}
if (waited)
cleanUpWorkQueue();
ArrayList<QueueEvent> heads = new ArrayList<>(blockedTasks.size() + MAX_PARALLEL_TASKS);
heads.addAll(blockedTasks.values());
// to clear out at least a few items in the queue before we read more items
if (heads.size() < MAX_BLOCKED_TASKS) {
//instead of reading MAX_PARALLEL_TASKS items always, we should only fetch as much as we can execute
int toFetch = Math.min(MAX_BLOCKED_TASKS - heads.size(), MAX_PARALLEL_TASKS - runningTasks.size());
List<QueueEvent> newTasks = workQueue.peekTopN(toFetch, excludedTasks, 2000L);
log.debug("Got {} tasks from work-queue : [{}]", newTasks.size(), newTasks);
heads.addAll(newTasks);
} else {
// Prevent free-spinning this loop.
Thread.sleep(1000);
}
if (isClosed)
break;
if (heads.isEmpty()) {
continue;
}
// clear it now; may get refilled below.
blockedTasks.clear();
taskBatch.batchId++;
boolean tooManyTasks = false;
for (QueueEvent head : heads) {
if (!tooManyTasks) {
synchronized (runningTasks) {
tooManyTasks = runningTasks.size() >= MAX_PARALLEL_TASKS;
}
}
if (tooManyTasks) {
// Too many tasks are running, just shove the rest into the "blocked" queue.
if (blockedTasks.size() < MAX_BLOCKED_TASKS)
blockedTasks.put(head.getId(), head);
continue;
}
if (runningZKTasks.contains(head.getId()))
continue;
final ZkNodeProps message = ZkNodeProps.load(head.getBytes());
OverseerMessageHandler messageHandler = selector.selectOverseerMessageHandler(message);
final String asyncId = message.getStr(ASYNC);
if (hasLeftOverItems) {
if (head.getId().equals(oldestItemInWorkQueue))
hasLeftOverItems = false;
if (asyncId != null && (completedMap.contains(asyncId) || failureMap.contains(asyncId))) {
log.debug("Found already processed task in workQueue, cleaning up. AsyncId [{}]", asyncId);
workQueue.remove(head);
continue;
}
}
String operation = message.getStr(Overseer.QUEUE_OPERATION);
OverseerMessageHandler.Lock lock = messageHandler.lockTask(message, taskBatch);
if (lock == null) {
log.debug("Exclusivity check failed for [{}]", message.toString());
//we may end crossing the size of the MAX_BLOCKED_TASKS. They are fine
if (blockedTasks.size() < MAX_BLOCKED_TASKS)
blockedTasks.put(head.getId(), head);
continue;
}
try {
markTaskAsRunning(head, asyncId);
log.debug("Marked task [{}] as running", head.getId());
} catch (KeeperException.NodeExistsException e) {
lock.unlock();
// This should never happen
log.error("Tried to pick up task [{}] when it was already running!", head.getId());
continue;
} catch (InterruptedException e) {
lock.unlock();
log.error("Thread interrupted while trying to pick task for execution.", head.getId());
Thread.currentThread().interrupt();
continue;
}
log.debug(messageHandler.getName() + ": Get the message id:" + head.getId() + " message:" + message.toString());
Runner runner = new Runner(messageHandler, message, operation, head, lock);
tpe.execute(runner);
}
} catch (KeeperException e) {
if (e.code() == KeeperException.Code.SESSIONEXPIRED) {
log.warn("Overseer cannot talk to ZK");
return;
}
SolrException.log(log, "", e);
} catch (InterruptedException e) {
Thread.currentThread().interrupt();
return;
} catch (Exception e) {
SolrException.log(log, "", e);
}
}
} finally {
this.close();
}
}
use of org.apache.solr.common.cloud.ZkNodeProps in project lucene-solr by apache.
the class OverseerTaskProcessor method amILeader.
protected LeaderStatus amILeader() {
String statsName = "collection_am_i_leader";
Timer.Context timerContext = stats.time(statsName);
boolean success = true;
try {
ZkNodeProps props = ZkNodeProps.load(zkStateReader.getZkClient().getData(Overseer.OVERSEER_ELECT + "/leader", null, null, true));
if (myId.equals(props.getStr(ID))) {
return LeaderStatus.YES;
}
} catch (KeeperException e) {
success = false;
if (e.code() == KeeperException.Code.CONNECTIONLOSS) {
log.error("", e);
return LeaderStatus.DONT_KNOW;
} else if (e.code() != KeeperException.Code.SESSIONEXPIRED) {
log.warn("", e);
}
} catch (InterruptedException e) {
success = false;
Thread.currentThread().interrupt();
} finally {
timerContext.stop();
if (success) {
stats.success(statsName);
} else {
stats.error(statsName);
}
}
log.info("According to ZK I (id=" + myId + ") am no longer a leader.");
return LeaderStatus.NO;
}
use of org.apache.solr.common.cloud.ZkNodeProps in project lucene-solr by apache.
the class OverseerTaskQueue method containsTaskWithRequestId.
/**
* Returns true if the queue contains a task with the specified async id.
*/
public boolean containsTaskWithRequestId(String requestIdKey, String requestId) throws KeeperException, InterruptedException {
List<String> childNames = zookeeper.getChildren(dir, null, true);
stats.setQueueLength(childNames.size());
for (String childName : childNames) {
if (childName != null && childName.startsWith(PREFIX)) {
try {
byte[] data = zookeeper.getData(dir + "/" + childName, null, null, true);
if (data != null) {
ZkNodeProps message = ZkNodeProps.load(data);
if (message.containsKey(requestIdKey)) {
LOG.debug(">>>> {}", message.get(requestIdKey));
if (message.get(requestIdKey).equals(requestId))
return true;
}
}
} catch (KeeperException.NoNodeException e) {
// Another client removed the node first, try next
}
}
}
return false;
}
use of org.apache.solr.common.cloud.ZkNodeProps in project lucene-solr by apache.
the class DeleteNodeCmd method cleanupReplicas.
static void cleanupReplicas(NamedList results, ClusterState clusterState, List<ZkNodeProps> sourceReplicas, OverseerCollectionMessageHandler ocmh, String node, String async) throws InterruptedException {
CountDownLatch cleanupLatch = new CountDownLatch(sourceReplicas.size());
for (ZkNodeProps sourceReplica : sourceReplicas) {
String coll = sourceReplica.getStr(COLLECTION_PROP);
String shard = sourceReplica.getStr(SHARD_ID_PROP);
log.info("Deleting replica for collection={} shard={} on node={}", coll, shard, node);
NamedList deleteResult = new NamedList();
try {
if (async != null)
sourceReplica = sourceReplica.plus(ASYNC, async);
((DeleteReplicaCmd) ocmh.commandMap.get(DELETEREPLICA)).deleteReplica(clusterState, sourceReplica.plus("parallel", "true"), deleteResult, () -> {
cleanupLatch.countDown();
if (deleteResult.get("failure") != null) {
synchronized (results) {
results.add("failure", String.format(Locale.ROOT, "Failed to delete replica for collection=%s shard=%s" + " on node=%s", coll, shard, node));
}
}
});
} catch (KeeperException e) {
log.warn("Error deleting ", e);
cleanupLatch.countDown();
} catch (Exception e) {
log.warn("Error deleting ", e);
cleanupLatch.countDown();
throw e;
}
}
log.debug("Waiting for delete node action to complete");
cleanupLatch.await(5, TimeUnit.MINUTES);
}
Aggregations