Exchange type topic throws exception
Closed this issue · 9 comments
When I use exchange type 'topic' while declaring exchange exception is thrown see below, but with type 'direct' everything works fine.
Exception:
[info] java.io.IOException:
[info] at com.rabbitmq.client.impl.AMQChannel.wrap(AMQChannel.java:106)
[info] at com.rabbitmq.client.impl.AMQChannel.wrap(AMQChannel.java:102)
[info] at com.rabbitmq.client.impl.AMQChannel.exnWrappingRpc(AMQChannel.java:124)
[info] at com.rabbitmq.client.impl.ChannelN.exchangeDeclare(ChannelN.java:703)
[info] at com.rabbitmq.client.impl.ChannelN.exchangeDeclare(ChannelN.java:672)
[info] at com.rabbitmq.client.impl.ChannelN.exchangeDeclare(ChannelN.java:720)
[info] at com.rabbitmq.client.impl.ChannelN.exchangeDeclare(ChannelN.java:61)
[info] ...
[info] Cause: com.rabbitmq.client.ShutdownSignalException: channel error; protocol method: #method<channel.close>(reply-code=406, reply-text=PRECONDITION_FAILED - inequivalent arg 'type' for exchange 'test-exchange' in vhost '/': received 'topic' but current is 'direct', class-id=40, method-id=10)Reporter completed abruptly with an exception after receiving event: TestFailed(Ordinal(0, 17),java.io.IOException was thrown.
An example recipe
val TestRMQNodeName = "rabbit-test"
val TestRMQNodePort = "5672"
val DecodedExchangeName = "test-exchange"
val config: EmbeddedRabbitMqConfig = new EmbeddedRabbitMqConfig.Builder()
.rabbitMqServerInitializationTimeoutInMillis(6000)
.defaultRabbitMqCtlTimeoutInMillis(6000)
.envVar(RabbitMqEnvVar.NODENAME, TestRMQNodeName)
.envVar(RabbitMqEnvVar.NODE_PORT, TestRMQNodePort)
.build()
val rabbitMq: EmbeddedRabbitMq = new EmbeddedRabbitMq(config)
rabbitMq.start()
val connFactory = new ConnectionFactory()
connFactory.setHost("localhost")
connFactory.setPort(5672)
connFactory.setVirtualHost("/")
connFactory.setUsername("guest")
connFactory.setPassword("guest")
val conn = connFactory.newConnection()
val channel = conn.createChannel()
channel.exchangeDeclare(DecodedExchangeName, "topic", true);
@noorul, i think the issue is specific to your setup.
As you can see, the error message says:
inequivalent arg 'type' for exchange 'test-exchange' in vhost '/': received 'topic' but current is 'direct', class-id=40, method-id=10)
They keywords here is inequivalent and received 'topic' but current is 'direct'.
This means that you're trying to re-declare an already-existing exchange which was declared previously with another "configuration".
RabbitMQ allows re-declarations to happen successfully IF the "configuration" matches what's already present.
The docs say:
Exchanges cannot be redeclared with different types. The client MUST not attempt to redeclare an existing exchange with a different type than used in the original Exchange.Declare method. Error code: not-allowed
That is what I suspected initially but unfortunately that is not the case. I could reproduce it by giving different names for different runs.
import io.arivera.oss.embedded.rabbitmq.EmbeddedRabbitMq;
import io.arivera.oss.embedded.rabbitmq.EmbeddedRabbitMqConfig;
import com.rabbitmq.client.Channel;
import com.rabbitmq.client.Connection;
import com.rabbitmq.client.ConnectionFactory;
import com.rabbitmq.client.GetResponse;
import org.junit.After;
import org.junit.Before;
import org.junit.Test;
import java.io.IOException;
import java.util.concurrent.TimeoutException;
import static org.hamcrest.CoreMatchers.equalTo;
import static org.junit.Assert.assertThat;
public class Issue32 {
private EmbeddedRabbitMq embeddedRabbitMq;
private EmbeddedRabbitMqConfig.Builder builder;
@Before
public void setUp() {
builder = new EmbeddedRabbitMqConfig.Builder();
embeddedRabbitMq = new EmbeddedRabbitMq(builder.build());
embeddedRabbitMq.start();
}
@After
public void tearDown() {
embeddedRabbitMq.stop();
}
@Test
public void name() throws IOException, TimeoutException {
ConnectionFactory connectionFactory = new ConnectionFactory();
connectionFactory.setHost("localhost");
connectionFactory.setPort(5672);
connectionFactory.setUsername("guest");
connectionFactory.setPassword("guest");
connectionFactory.setVirtualHost("/");
Connection connection = connectionFactory.newConnection();
Channel channel = null;
try {
channel = connection.createChannel();
channel.exchangeDeclare("my-exchange", "topic", true, false, null);
channel.queueDeclare("my-queue", true, false, false, null);
channel.queueBind("my-queue", "my-exchange", "foo.*", null);
String messageToSend = "hello world!";
channel.basicPublish("my-exchange", "foo.bar", null, messageToSend.getBytes("utf-8"));
GetResponse getResponse = channel.basicGet("my-queue", true);
String messageReceived = new String(getResponse.getBody(), "utf-8");
assertThat(messageReceived, equalTo(messageToSend));
System.out.println("Message received: " + messageReceived);
} finally {
if (channel != null) { channel.close(); }
connection.close();
}
}
}
gives me:
20:17:05.865 DEBUG main i.a.o.e.r.E.P.rabbitmq-server - Process started.
20:17:06.848 INFO Thread-2 i.a.o.e.r.E.P.rabbitmq-server -
20:17:06.848 INFO Thread-2 i.a.o.e.r.E.P.rabbitmq-server - RabbitMQ 3.6.6. Copyright (C) 2007-2016 Pivotal Software, Inc.
20:17:06.848 INFO Thread-2 i.a.o.e.r.E.P.rabbitmq-server - ## ## Licensed under the MPL. See http://www.rabbitmq.com/
20:17:06.848 INFO Thread-2 i.a.o.e.r.E.P.rabbitmq-server - ## ##
20:17:06.848 INFO Thread-2 i.a.o.e.r.E.P.rabbitmq-server - ########## Logs: /var/folders/5v/x81_dvcn11ddrkb2ybs56k_r0000gn/T/rabbitmq_server-3.6.6/var/log/rabbitmq/rabbit@rivera-mbp.log
20:17:06.848 INFO Thread-2 i.a.o.e.r.E.P.rabbitmq-server - ###### ## /var/folders/5v/x81_dvcn11ddrkb2ybs56k_r0000gn/T/rabbitmq_server-3.6.6/var/log/rabbitmq/rabbit@rivera-mbp-sasl.log
20:17:06.848 INFO Thread-2 i.a.o.e.r.E.P.rabbitmq-server - ##########
20:17:06.848 INFO Thread-2 i.a.o.e.r.E.P.rabbitmq-server - Starting broker...
20:17:07.449 INFO Thread-2 i.a.o.e.r.E.P.rabbitmq-server - completed with 0 plugins.
Message received: hello world!
It looks like stop() leaves epmd daemon running.
$ ps -ef | grep -i [r]abbit
501 85037 1 0 8:34PM ?? 0:00.00 /var/folders/j6/4qlzsrl541q6nqyfhbkmfnvr0000gn/T/rabbitmq_server-3.6.6/erts-6.4/bin/../../erts-6.4/bin/epmd -daemon
$
If I kill it and run it is working.
stop() will invoke the command to shut down RabbitMQ and wait for completion of the command.
if something goes wrong and the JVM exits, the process might continue to run.
In some systems, this is a bigger problem than in others. For example, in Mac OS, the process always seems to be shut-down (perhaps because it's considered a child process of the JVM), but in Ubuntu, it doesn't.
See PR #28 for the same report and test case that proves this problem.
I've not been able to find a consistent/good solution to enforce the process to always be killed, but it seems to be a common issue with all libraries that initiate external processes (EmbeddedRedis, EmbeddedMongo, etc.)
If you do find a solution that works, a PR will be most welcomed
It looks like rabbitmq also persists metadata.
/var/folders/j6/4qlzsrl541q6nqyfhbkmfnvr0000gn/T/rabbitmq_server-3.6.6/var/lib/rabbitmq/mnesia/rabbit-test
$ ls -la
total 184
drwxr-xr-x 20 nislam staff 680 Apr 8 20:46 .
drwxr-xr-x 5 nislam staff 170 Apr 8 20:46 ..
-rw-r--r-- 1 nislam staff 171 Apr 8 20:46 DECISION_TAB.LOG
-rw-r--r-- 1 nislam staff 718 Apr 8 20:46 LATEST.LOG
-rw-r--r-- 1 nislam staff 75 Apr 8 20:46 cluster_nodes.config
drwxr-xr-x 6 nislam staff 204 Apr 8 20:46 msg_store_persistent
drwxr-xr-x 6 nislam staff 204 Apr 8 20:46 msg_store_transient
-rw-r--r-- 1 nislam staff 37 Apr 8 20:46 nodes_running_at_shutdown
-rw-r--r-- 1 nislam staff 1335 Apr 6 12:41 rabbit_durable_exchange.DCD
-rw-r--r-- 1 nislam staff 591 Apr 8 20:34 rabbit_durable_exchange.DCL
-rw-r--r-- 1 nislam staff 285 Apr 8 20:46 rabbit_durable_queue.DCD
-rw-r--r-- 1 nislam staff 451 Apr 8 20:34 rabbit_durable_route.DCD
-rw-r--r-- 1 nislam staff 191 Apr 6 12:41 rabbit_runtime_parameters.DCD
-rw-r--r-- 1 nislam staff 4 Apr 8 20:46 rabbit_serial
-rw-r--r-- 1 nislam staff 241 Apr 6 12:41 rabbit_user.DCD
-rw-r--r-- 1 nislam staff 204 Apr 6 12:41 rabbit_user_permission.DCD
-rw-r--r-- 1 nislam staff 145 Apr 6 12:41 rabbit_vhost.DCD
-rw-r--r-- 1 nislam staff 5965 Apr 8 20:46 recovery.dets
-rw-r--r-- 1 nislam staff 24767 Apr 6 12:41 schema.DAT
-rw-r--r-- 1 nislam staff 283 Apr 6 12:32 schema_version
I think I can close this issue now. I never knew that rabbitmq actually persists metadata.
But do you think I should open a ticket for epmd daemon being running even after calling stop() ?
Oops! I did not read you previous comment regarding stop() before sending out my last comment. So you can ignore that. Thank you for your time.
No worries. Happy coding :)
One of the reason could be that you have not specified the consumer type to direct and by default, it takes topic