com.datastax.driver.core.ControlConnection Maven / Gradle / Ivy
Go to download
Show more of this group Show more artifacts with this name
Show all versions of cassandra-driver-core Show documentation
Show all versions of cassandra-driver-core Show documentation
A driver for Apache Cassandra 1.2+ that works exclusively with the Cassandra Query Language version 3
(CQL3) and Cassandra's binary protocol.
/*
* Copyright (C) 2012-2015 DataStax Inc.
*
* Licensed 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 com.datastax.driver.core;
import com.datastax.driver.core.exceptions.*;
import com.google.common.annotations.VisibleForTesting;
import com.google.common.base.Objects;
import com.google.common.util.concurrent.ListenableFuture;
import org.slf4j.Logger;
import org.slf4j.LoggerFactory;
import java.net.InetAddress;
import java.net.InetSocketAddress;
import java.net.UnknownHostException;
import java.util.*;
import java.util.concurrent.ExecutionException;
import java.util.concurrent.TimeUnit;
import java.util.concurrent.atomic.AtomicReference;
import static com.datastax.driver.core.SchemaElement.KEYSPACE;
class ControlConnection implements Connection.Owner {
private static final Logger logger = LoggerFactory.getLogger(ControlConnection.class);
private static final boolean EXTENDED_PEER_CHECK = SystemProperties.getBoolean("com.datastax.driver.EXTENDED_PEER_CHECK", true);
private static final InetAddress bindAllAddress;
static {
try {
bindAllAddress = InetAddress.getByAddress(new byte[4]);
} catch (UnknownHostException e) {
throw new RuntimeException(e);
}
}
private static final String SELECT_PEERS = "SELECT * FROM system.peers";
private static final String SELECT_LOCAL = "SELECT * FROM system.local WHERE key='local'";
private static final String SELECT_SCHEMA_PEERS = "SELECT peer, rpc_address, schema_version FROM system.peers";
private static final String SELECT_SCHEMA_LOCAL = "SELECT schema_version FROM system.local WHERE key='local'";
@VisibleForTesting
final AtomicReference connectionRef = new AtomicReference();
private final Cluster.Manager cluster;
private final AtomicReference> reconnectionAttempt = new AtomicReference>();
private volatile boolean isShutdown;
public ControlConnection(Cluster.Manager manager) {
this.cluster = manager;
}
// Only for the initial connection. Does not schedule retries if it fails
void connect() throws UnsupportedProtocolVersionException {
if (isShutdown)
return;
// NB: at this stage, allHosts() only contains the initial contact points
List hosts = new ArrayList(cluster.metadata.allHosts());
// shuffle so that multiple clients with the same contact points don't all pick the same control host
Collections.shuffle(hosts);
setNewConnection(reconnectInternal(hosts.iterator(), true));
}
CloseFuture closeAsync() {
// We don't have to be fancy here. We just set a flag so that we stop trying to reconnect (and thus change the
// connection used) and shutdown the current one.
isShutdown = true;
// Cancel any reconnection attempt in progress
ListenableFuture r = reconnectionAttempt.get();
if (r != null)
r.cancel(false);
Connection connection = connectionRef.get();
return connection == null ? CloseFuture.immediateFuture() : connection.closeAsync().force();
}
Host connectedHost() {
Connection current = connectionRef.get();
return (current == null)
? null
: cluster.metadata.getHost(current.address);
}
void triggerReconnect() {
backgroundReconnect(0);
}
/**
* @param initialDelayMs if >=0, bypass the schedule and use this for the first call
*/
private void backgroundReconnect(long initialDelayMs) {
if (isShutdown)
return;
// Abort if a reconnection is already in progress. This is not thread-safe: two threads might race through this and both
// schedule a reconnection; in that case AbstractReconnectionHandler knows how to deal with it correctly.
// But this cheap check can help us avoid creating the object unnecessarily.
ListenableFuture reconnection = reconnectionAttempt.get();
if (reconnection != null && !reconnection.isDone())
return;
new AbstractReconnectionHandler("Control connection", cluster.reconnectionExecutor, cluster.reconnectionPolicy().newSchedule(), reconnectionAttempt, initialDelayMs) {
@Override
protected Connection tryReconnect() throws ConnectionException {
if (isShutdown)
throw new ConnectionException(null, "Control connection was shut down");
try {
return reconnectInternal(queryPlan(), false);
} catch (NoHostAvailableException e) {
throw new ConnectionException(null, e.getMessage());
} catch (UnsupportedProtocolVersionException e) {
// reconnectInternal only propagate those if we've not decided on the protocol version yet,
// which should only happen on the initial connection and thus in connect() but never here.
throw new AssertionError();
}
}
@Override
protected void onReconnection(Connection connection) {
if (isShutdown) {
connection.closeAsync().force();
return;
}
setNewConnection(connection);
}
@Override
protected boolean onConnectionException(ConnectionException e, long nextDelayMs) {
if (isShutdown)
return false;
logger.error("[Control connection] Cannot connect to any host, scheduling retry in {} milliseconds", nextDelayMs);
return true;
}
@Override
protected boolean onUnknownException(Exception e, long nextDelayMs) {
if (isShutdown)
return false;
logger.error(String.format("[Control connection] Unknown error during reconnection, scheduling retry in %d milliseconds", nextDelayMs), e);
return true;
}
}.start();
}
private Iterator queryPlan() {
return cluster.loadBalancingPolicy().newQueryPlan(null, Statement.DEFAULT);
}
private void signalError() {
Connection connection = connectionRef.get();
if (connection != null)
connection.closeAsync().force();
// If the error caused the host to go down, onDown might have already triggered a reconnect.
// But backgroundReconnect knows how to deal with that.
backgroundReconnect(0);
}
private void setNewConnection(Connection newConnection) {
Host.statesLogger.debug("[Control connection] established to {}", newConnection.address);
newConnection.setOwner(this);
Connection old = connectionRef.getAndSet(newConnection);
if (old != null && !old.isClosed())
old.closeAsync().force();
}
private Connection reconnectInternal(Iterator iter, boolean isInitialConnection) throws UnsupportedProtocolVersionException {
Map errors = null;
Host host = null;
try {
while (iter.hasNext()) {
host = iter.next();
if (!host.convictionPolicy.canReconnectNow())
continue;
try {
return tryConnect(host, isInitialConnection);
} catch (ConnectionException e) {
errors = logError(host, e, errors, iter);
if (isInitialConnection) {
// Mark the host down right away so that we don't try it again during the initialization process.
// We don't call cluster.triggerOnDown because it does a bunch of other things we don't want to do here (notify LBP, etc.)
host.setDown();
}
} catch (ExecutionException e) {
errors = logError(host, e.getCause(), errors, iter);
} catch (UnsupportedProtocolVersionException e) {
// If it's the very first node we've connected to, rethrow the exception and
// Cluster.init() will handle it. Otherwise, just mark this node in error.
if (cluster.protocolVersion() == null)
throw e;
logger.debug("Ignoring host {}: {}", host, e.getMessage());
errors = logError(host, e, errors, iter);
} catch (ClusterNameMismatchException e) {
logger.debug("Ignoring host {}: {}", host, e.getMessage());
errors = logError(host, e, errors, iter);
}
}
} catch (InterruptedException e) {
// Sets interrupted status
Thread.currentThread().interrupt();
// Indicates that all remaining hosts are skipped due to the interruption
if (host != null)
errors = logError(host, new DriverException("Connection thread interrupted"), errors, iter);
while (iter.hasNext())
errors = logError(iter.next(), new DriverException("Connection thread interrupted"), errors, iter);
}
throw new NoHostAvailableException(errors == null ? Collections.emptyMap() : errors);
}
private static Map logError(Host host, Throwable exception, Map errors, Iterator iter) {
if (errors == null)
errors = new HashMap();
errors.put(host.getSocketAddress(), exception);
if (logger.isDebugEnabled()) {
if (iter.hasNext()) {
logger.debug(String.format("[Control connection] error on %s connection, trying next host", host), exception);
} else {
logger.debug(String.format("[Control connection] error on %s connection, no more host to try", host), exception);
}
}
return errors;
}
private Connection tryConnect(Host host, boolean isInitialConnection) throws ConnectionException, ExecutionException, InterruptedException, UnsupportedProtocolVersionException, ClusterNameMismatchException {
Connection connection = cluster.connectionFactory.open(host);
// If no protocol version was specified, set the default as soon as a connection succeeds (it's needed to parse UDTs in refreshSchema)
if (cluster.connectionFactory.protocolVersion == null)
cluster.connectionFactory.protocolVersion = ProtocolVersion.NEWEST_SUPPORTED;
try {
logger.trace("[Control connection] Registering for events");
List evs = Arrays.asList(
ProtocolEvent.Type.TOPOLOGY_CHANGE,
ProtocolEvent.Type.STATUS_CHANGE,
ProtocolEvent.Type.SCHEMA_CHANGE
);
connection.write(new Requests.Register(evs));
// We need to refresh the node list first so we know about the cassandra version of
// the node we're connecting to.
// This will create the token map for the first time, but it will be incomplete
// due to the lack of keyspace information
refreshNodeListAndTokenMap(connection, cluster, isInitialConnection, true);
// refresh schema will also update the token map again,
// this time with information about keyspaces
logger.debug("[Control connection] Refreshing schema");
refreshSchema(connection, null, null, null, null, cluster);
return connection;
} catch (BusyConnectionException e) {
connection.closeAsync().force();
throw new DriverInternalError("Newly created connection should not be busy");
} catch (InterruptedException e) {
connection.closeAsync().force();
throw e;
} catch (ConnectionException e) {
connection.closeAsync().force();
throw e;
} catch (ExecutionException e) {
connection.closeAsync().force();
throw e;
} catch (RuntimeException e) {
connection.closeAsync().force();
throw e;
}
}
public void refreshSchema(SchemaElement targetType, String targetKeyspace, String targetName, List signature) throws InterruptedException {
logger.debug("[Control connection] Refreshing schema for {}{}",
targetType == null ? "everything" : targetKeyspace,
(targetType == KEYSPACE) ? "" : "." + targetName + " (" + targetType + ")");
try {
Connection c = connectionRef.get();
// At startup, when we add the initial nodes, this will be null, which is ok
if (c == null || c.isClosed())
return;
refreshSchema(c, targetType, targetKeyspace, targetName, signature, cluster);
} catch (ConnectionException e) {
logger.debug("[Control connection] Connection error while refreshing schema ({})", e.getMessage());
signalError();
} catch (ExecutionException e) {
// If we're being shutdown during schema refresh, this can happen. That's fine so don't scare the user.
if (!isShutdown)
logger.error("[Control connection] Unexpected error while refreshing schema", e);
signalError();
} catch (BusyConnectionException e) {
logger.debug("[Control connection] Connection is busy, reconnecting");
signalError();
}
}
static void refreshSchema(Connection connection, SchemaElement targetType, String targetKeyspace, String targetName, List targetSignature, Cluster.Manager cluster) throws ConnectionException, BusyConnectionException, ExecutionException, InterruptedException {
Host host = cluster.metadata.getHost(connection.address);
// Neither host, nor it's version should be null. But instead of dying if there is a race or something, we can kind of try to infer
// a Cassandra version from the protocol version (this is not full proof, we can have the protocol 1 against C* 2.0+, but it's worth
// a shot, and since we log in this case, it should be relatively easy to debug when if this ever fail).
VersionNumber cassandraVersion;
if (host == null || host.getCassandraVersion() == null) {
cassandraVersion = cluster.protocolVersion().minCassandraVersion();
logger.warn("Cannot find Cassandra version for host {} to parse the schema, using {} based on protocol version in use. "
+ "If parsing the schema fails, this could be the cause", connection.address, cassandraVersion);
} else {
cassandraVersion = host.getCassandraVersion();
}
SchemaParser.forVersion(cassandraVersion)
.refresh(cluster.getCluster(),
targetType, targetKeyspace, targetName, targetSignature,
connection, cassandraVersion);
}
void refreshNodeListAndTokenMap() {
Connection c = connectionRef.get();
// At startup, when we add the initial nodes, this will be null, which is ok
if (c == null || c.isClosed())
return;
try {
refreshNodeListAndTokenMap(c, cluster, false, true);
} catch (ConnectionException e) {
logger.debug("[Control connection] Connection error while refreshing node list and token map ({})", e.getMessage());
signalError();
} catch (ExecutionException e) {
// If we're being shutdown during refresh, this can happen. That's fine so don't scare the user.
if (!isShutdown)
logger.error("[Control connection] Unexpected error while refreshing node list and token map", e);
signalError();
} catch (BusyConnectionException e) {
logger.debug("[Control connection] Connection is busy, reconnecting");
signalError();
} catch (InterruptedException e) {
Thread.currentThread().interrupt();
logger.debug("[Control connection] Interrupted while refreshing node list and token map, skipping it.");
}
}
private static InetSocketAddress rpcAddressForPeerHost(Row peersRow, InetSocketAddress connectedHost, Cluster.Manager cluster) {
// after CASSANDRA-9436, system.peers contains the following inet columns:
// - peer: this is actually broadcast_address
// - rpc_address: the address we are looking for (this corresponds to broadcast_rpc_address in the peer's cassandra yaml file;
// if this setting if unset, it defaults to the value for rpc_address or rpc_interface)
// - preferred_ip: used by Ec2MultiRegionSnitch and GossipingPropertyFileSnitch, possibly others; contents unclear
InetAddress broadcastAddress = peersRow.getInet("peer");
InetAddress rpcAddress = peersRow.getInet("rpc_address");
if (broadcastAddress == null) {
return null;
} else if (broadcastAddress.equals(connectedHost.getAddress()) || (rpcAddress != null && rpcAddress.equals(connectedHost.getAddress()))) {
// Some DSE versions were inserting a line for the local node in peers (with mostly null values). This has been fixed, but if we
// detect that's the case, ignore it as it's not really a big deal.
logger.debug("System.peers on node {} has a line for itself. This is not normal but is a known problem of some DSE version. Ignoring the entry.", connectedHost);
return null;
} else if (rpcAddress == null) {
return null;
} else if (rpcAddress.equals(bindAllAddress)) {
logger.warn("Found host with 0.0.0.0 as rpc_address, using broadcast_address ({}) to contact it instead. If this is incorrect you should avoid the use of 0.0.0.0 server side.", broadcastAddress);
rpcAddress = broadcastAddress;
}
return cluster.translateAddress(rpcAddress);
}
private Row fetchNodeInfo(Host host, Connection c) throws ConnectionException, BusyConnectionException, ExecutionException, InterruptedException {
boolean isConnectedHost = c.address.equals(host.getSocketAddress());
if (isConnectedHost || host.getBroadcastAddress() != null) {
DefaultResultSetFuture future = isConnectedHost
? new DefaultResultSetFuture(null, cluster.protocolVersion(), new Requests.Query(SELECT_LOCAL))
: new DefaultResultSetFuture(null, cluster.protocolVersion(), new Requests.Query(SELECT_PEERS + " WHERE peer='" + host.getBroadcastAddress().getHostAddress() + '\''));
c.write(future);
Row row = future.get().one();
if (row != null) {
return row;
} else {
logger.debug("Could not find peer with broadcast address {}, " +
"falling back to a full system.peers scan to fetch info for {} " +
"(this can happen if the broadcast address changed)", host.getBroadcastAddress(), host);
}
}
// We have to fetch the whole peers table and find the host we're looking for
DefaultResultSetFuture future = new DefaultResultSetFuture(null, cluster.protocolVersion(), new Requests.Query(SELECT_PEERS));
c.write(future);
for (Row row : future.get()) {
InetSocketAddress addr = rpcAddressForPeerHost(row, c.address, cluster);
if (addr != null && addr.equals(host.getSocketAddress()))
return row;
}
return null;
}
/**
* @return whether we have enough information to bring the node back up
*/
boolean refreshNodeInfo(Host host) {
Connection c = connectionRef.get();
// At startup, when we add the initial nodes, this will be null, which is ok
if (c == null || c.isClosed())
return true;
logger.debug("[Control connection] Refreshing node info on {}", host);
try {
Row row = fetchNodeInfo(host, c);
if (row == null) {
if (c.isDefunct()) {
logger.debug("Control connection is down, could not refresh node info");
// Keep going with what we currently know about the node, otherwise we will ignore all nodes
// until the control connection is back up (which leads to a catch-22 if there is only one)
return true;
} else {
logger.warn("No row found for host {} in {}'s peers system table. {} will be ignored.", host.getAddress(), c.address, host.getAddress());
return false;
}
// Ignore hosts with a null rpc_address, as this is most likely a phantom row in system.peers (JAVA-428).
// Don't test this for the control host since we're already connected to it anyway, and we read the info from system.local
// which didn't have an rpc_address column (JAVA-546) until CASSANDRA-9436
} else if (!c.address.equals(host.getSocketAddress()) && !isValidPeer(row, true)) {
return false;
}
updateInfo(host, row, cluster, false);
return true;
} catch (ConnectionException e) {
logger.debug("[Control connection] Connection error while refreshing node info ({})", e.getMessage());
signalError();
} catch (ExecutionException e) {
// If we're being shutdown during refresh, this can happen. That's fine so don't scare the user.
if (!isShutdown)
logger.debug("[Control connection] Unexpected error while refreshing node info", e);
signalError();
} catch (BusyConnectionException e) {
logger.debug("[Control connection] Connection is busy, reconnecting");
signalError();
} catch (InterruptedException e) {
Thread.currentThread().interrupt();
logger.debug("[Control connection] Interrupted while refreshing node info, skipping it.");
} catch (Exception e) {
logger.debug("[Control connection] Unexpected error while refreshing node info", e);
signalError();
}
// If we got an exception, always return true. Otherwise a faulty control connection would cause
// reconnected hosts to be ignored permanently.
return true;
}
// row can come either from the 'local' table or the 'peers' one
private static void updateInfo(Host host, Row row, Cluster.Manager cluster, boolean isInitialConnection) {
if (!row.isNull("data_center") || !row.isNull("rack"))
updateLocationInfo(host, row.getString("data_center"), row.getString("rack"), isInitialConnection, cluster);
String version = row.getString("release_version");
host.setVersion(version);
// Before CASSANDRA-9436 local row did not contain any info about the host addresses.
// After CASSANDRA-9436 (2.0.16, 2.1.6, 2.2.0 rc1) local row contains two new columns:
// - broadcast_address
// - rpc_address
// After CASSANDRA-9603 (2.0.17, 2.1.8, 2.2.0 rc2) local row contains one more column:
// - listen_address
InetAddress broadcastAddress = null;
if (row.getColumnDefinitions().contains("peer")) { // system.peers
broadcastAddress = row.getInet("peer");
} else if (row.getColumnDefinitions().contains("broadcast_address")) { // system.local
broadcastAddress = row.getInet("broadcast_address");
}
host.setBroadcastAddress(broadcastAddress);
// in system.local only for C* versions >= 2.0.17, 2.1.8, 2.2.0 rc2,
// not yet in system.peers as of C* 3.2
InetAddress listenAddress = row.getColumnDefinitions().contains("listen_address")
? row.getInet("listen_address")
: null;
host.setListenAddress(listenAddress);
if (row.getColumnDefinitions().contains("workload")) {
String dseWorkload = row.getString("workload");
host.setDseWorkload(dseWorkload);
}
if (row.getColumnDefinitions().contains("graph")) {
boolean isDseGraph = row.getBool("graph");
host.setDseGraphEnabled(isDseGraph);
}
if (row.getColumnDefinitions().contains("dse_version")) {
String dseVersion = row.getString("dse_version");
host.setDseVersion(dseVersion);
}
}
private static void updateLocationInfo(Host host, String datacenter, String rack, boolean isInitialConnection, Cluster.Manager cluster) {
if (Objects.equal(host.getDatacenter(), datacenter) && Objects.equal(host.getRack(), rack))
return;
// If the dc/rack information changes for an existing node, we need to update the load balancing policy.
// For that, we remove and re-add the node against the policy. Not the most elegant, and assumes
// that the policy will update correctly, but in practice this should work.
if (!isInitialConnection)
cluster.loadBalancingPolicy().onDown(host);
host.setLocationInfo(datacenter, rack);
if (!isInitialConnection)
cluster.loadBalancingPolicy().onAdd(host);
}
private static void refreshNodeListAndTokenMap(Connection connection, Cluster.Manager cluster, boolean isInitialConnection, boolean logInvalidPeers) throws ConnectionException, BusyConnectionException, ExecutionException, InterruptedException {
logger.debug("[Control connection] Refreshing node list and token map");
boolean metadataEnabled = cluster.configuration.getQueryOptions().isMetadataEnabled();
// Make sure we're up to date on nodes and tokens
DefaultResultSetFuture localFuture = new DefaultResultSetFuture(null, cluster.protocolVersion(), new Requests.Query(SELECT_LOCAL));
DefaultResultSetFuture peersFuture = new DefaultResultSetFuture(null, cluster.protocolVersion(), new Requests.Query(SELECT_PEERS));
connection.write(localFuture);
connection.write(peersFuture);
String partitioner = null;
Token.Factory factory = null;
Map> tokenMap = new HashMap>();
// Update cluster name, DC and rack for the one node we are connected to
Row localRow = localFuture.get().one();
if (localRow != null) {
String clusterName = localRow.getString("cluster_name");
if (clusterName != null)
cluster.metadata.clusterName = clusterName;
partitioner = localRow.getString("partitioner");
if (partitioner != null) {
cluster.metadata.partitioner = partitioner;
factory = Token.getFactory(partitioner);
}
Host host = cluster.metadata.getHost(connection.address);
// In theory host can't be null. However there is no point in risking a NPE in case we
// have a race between a node removal and this.
if (host == null) {
logger.debug("Host in local system table ({}) unknown to us (ok if said host just got removed)", connection.address);
} else {
updateInfo(host, localRow, cluster, isInitialConnection);
if (metadataEnabled && factory != null) {
Set tokensStr = localRow.getSet("tokens", String.class);
if (!tokensStr.isEmpty()) {
Set tokens = toTokens(factory, tokensStr);
tokenMap.put(host, tokens);
}
}
}
}
List foundHosts = new ArrayList();
List dcs = new ArrayList();
List racks = new ArrayList();
List cassandraVersions = new ArrayList();
List broadcastAddresses = new ArrayList();
List listenAddresses = new ArrayList();
List> allTokens = new ArrayList>();
List dseVersions = new ArrayList();
List dseGraphEnabled = new ArrayList();
List dseWorkloads = new ArrayList();
for (Row row : peersFuture.get()) {
if (!isValidPeer(row, logInvalidPeers))
continue;
InetSocketAddress rpcAddress = rpcAddressForPeerHost(row, connection.address, cluster);
if (rpcAddress == null)
continue;
foundHosts.add(rpcAddress);
dcs.add(row.getString("data_center"));
racks.add(row.getString("rack"));
cassandraVersions.add(row.getString("release_version"));
broadcastAddresses.add(row.getInet("peer"));
if (metadataEnabled && factory != null) {
Set tokensStr = row.getSet("tokens", String.class);
Set tokens = null;
if (!tokensStr.isEmpty()) {
tokens = toTokens(factory, tokensStr);
}
allTokens.add(tokens);
}
InetAddress listenAddress = row.getColumnDefinitions().contains("listen_address") ? row.getInet("listen_address") : null;
listenAddresses.add(listenAddress);
String dseWorkload = row.getColumnDefinitions().contains("workload") ? row.getString("workload") : null;
dseWorkloads.add(dseWorkload);
Boolean isDseGraph = row.getColumnDefinitions().contains("graph") ? row.getBool("graph") : null;
dseGraphEnabled.add(isDseGraph);
String dseVersion = row.getColumnDefinitions().contains("dse_version") ? row.getString("dse_version") : null;
dseVersions.add(dseVersion);
}
for (int i = 0; i < foundHosts.size(); i++) {
Host host = cluster.metadata.getHost(foundHosts.get(i));
boolean isNew = false;
if (host == null) {
// We don't know that node, create the Host object but wait until we've set the known
// info before signaling the addition.
Host newHost = cluster.metadata.newHost(foundHosts.get(i));
Host existing = cluster.metadata.addIfAbsent(newHost);
if (existing == null) {
host = newHost;
isNew = true;
} else {
host = existing;
isNew = false;
}
}
if (dcs.get(i) != null || racks.get(i) != null)
updateLocationInfo(host, dcs.get(i), racks.get(i), isInitialConnection, cluster);
if (cassandraVersions.get(i) != null)
host.setVersion(cassandraVersions.get(i));
if (broadcastAddresses.get(i) != null)
host.setBroadcastAddress(broadcastAddresses.get(i));
if (listenAddresses.get(i) != null)
host.setListenAddress(listenAddresses.get(i));
if (dseVersions.get(i) != null)
host.setDseVersion(dseVersions.get(i));
if (dseWorkloads.get(i) != null)
host.setDseWorkload(dseWorkloads.get(i));
if (dseGraphEnabled.get(i) != null)
host.setDseGraphEnabled(dseGraphEnabled.get(i));
if (metadataEnabled && factory != null && allTokens.get(i) != null)
tokenMap.put(host, allTokens.get(i));
if (isNew && !isInitialConnection)
cluster.triggerOnAdd(host);
}
// Removes all those that seems to have been removed (since we lost the control connection)
Set foundHostsSet = new HashSet(foundHosts);
for (Host host : cluster.metadata.allHosts())
if (!host.getSocketAddress().equals(connection.address) && !foundHostsSet.contains(host.getSocketAddress()))
cluster.removeHost(host, isInitialConnection);
if (metadataEnabled && factory != null && !tokenMap.isEmpty())
cluster.metadata.rebuildTokenMap(factory, tokenMap);
}
private static Set toTokens(Token.Factory factory, Set tokensStr) {
Set tokens = new LinkedHashSet(tokensStr.size());
for (String tokenStr : tokensStr) {
tokens.add(factory.fromString(tokenStr));
}
return tokens;
}
private static boolean isValidPeer(Row peerRow, boolean logIfInvalid) {
boolean isValid = peerRow.getColumnDefinitions().contains("rpc_address")
&& !peerRow.isNull("rpc_address");
if (EXTENDED_PEER_CHECK) {
isValid &= peerRow.getColumnDefinitions().contains("host_id")
&& !peerRow.isNull("host_id")
&& peerRow.getColumnDefinitions().contains("data_center")
&& !peerRow.isNull("data_center")
&& peerRow.getColumnDefinitions().contains("rack")
&& !peerRow.isNull("rack")
&& peerRow.getColumnDefinitions().contains("tokens")
&& !peerRow.isNull("tokens");
}
if (!isValid && logIfInvalid)
logger.warn("Found invalid row in system.peers: {}. " +
"This is likely a gossip or snitch issue, this host will be ignored.", formatInvalidPeer(peerRow));
return isValid;
}
// Custom formatting to avoid spamming the logs if 'tokens' is present and contains a gazillion tokens
private static String formatInvalidPeer(Row peerRow) {
StringBuilder sb = new StringBuilder("[peer=" + peerRow.getInet("peer"));
formatMissingOrNullColumn(peerRow, "rpc_address", sb);
if (EXTENDED_PEER_CHECK) {
formatMissingOrNullColumn(peerRow, "host_id", sb);
formatMissingOrNullColumn(peerRow, "data_center", sb);
formatMissingOrNullColumn(peerRow, "rack", sb);
formatMissingOrNullColumn(peerRow, "tokens", sb);
}
sb.append("]");
return sb.toString();
}
private static void formatMissingOrNullColumn(Row peerRow, String columnName, StringBuilder sb) {
if (!peerRow.getColumnDefinitions().contains(columnName))
sb.append(", missing ").append(columnName);
else if (peerRow.isNull(columnName))
sb.append(", ").append(columnName).append("=null");
}
static boolean waitForSchemaAgreement(Connection connection, Cluster.Manager cluster) throws ConnectionException, BusyConnectionException, ExecutionException, InterruptedException {
long start = System.nanoTime();
long elapsed = 0;
int maxSchemaAgreementWaitSeconds = cluster.configuration.getProtocolOptions().getMaxSchemaAgreementWaitSeconds();
while (elapsed < maxSchemaAgreementWaitSeconds * 1000) {
if (checkSchemaAgreement(connection, cluster))
return true;
// let's not flood the node too much
Thread.sleep(200);
elapsed = Cluster.timeSince(start, TimeUnit.MILLISECONDS);
}
return false;
}
private static boolean checkSchemaAgreement(Connection connection, Cluster.Manager cluster) throws InterruptedException, ExecutionException {
DefaultResultSetFuture peersFuture = new DefaultResultSetFuture(null, cluster.protocolVersion(), new Requests.Query(SELECT_SCHEMA_PEERS));
DefaultResultSetFuture localFuture = new DefaultResultSetFuture(null, cluster.protocolVersion(), new Requests.Query(SELECT_SCHEMA_LOCAL));
connection.write(peersFuture);
connection.write(localFuture);
Set versions = new HashSet();
Row localRow = localFuture.get().one();
if (localRow != null && !localRow.isNull("schema_version"))
versions.add(localRow.getUUID("schema_version"));
for (Row row : peersFuture.get()) {
InetSocketAddress addr = rpcAddressForPeerHost(row, connection.address, cluster);
if (addr == null || row.isNull("schema_version"))
continue;
Host peer = cluster.metadata.getHost(addr);
if (peer != null && peer.isUp())
versions.add(row.getUUID("schema_version"));
}
logger.debug("Checking for schema agreement: versions are {}", versions);
return versions.size() <= 1;
}
boolean checkSchemaAgreement() throws ConnectionException, BusyConnectionException, InterruptedException, ExecutionException {
Connection connection = connectionRef.get();
return connection != null &&
!connection.isClosed() &&
checkSchemaAgreement(connection, cluster);
}
boolean isOpen() {
Connection c = connectionRef.get();
return c != null && !c.isClosed();
}
public void onUp(Host host) {
}
public void onAdd(Host host) {
}
public void onDown(Host host) {
onHostGone(host);
}
public void onRemove(Host host) {
onHostGone(host);
}
private void onHostGone(Host host) {
Connection current = connectionRef.get();
if (current != null && current.address.equals(host.getSocketAddress())) {
logger.debug("[Control connection] {} is down/removed and it was the control host, triggering reconnect",
current.address);
if (!current.isClosed())
current.closeAsync().force();
backgroundReconnect(0);
}
}
@Override
public void onConnectionDefunct(Connection connection) {
if (connection == connectionRef.get())
backgroundReconnect(0);
}
}