No way to run ping pong

[Paulo Suzart]

Hi,

It can be my machines (a mac and a ubuntu both with latest java 7), but it is not possible to run the examples from pulsar.

Simply cloning the project and switching to 0.5.1 and running:

lein with-profile cluster update-in :jvm-opts conj '"-Dgalaxy.nodeId=2"' '"-Dgalaxy.port=7052"' '"-Dgalaxy.slave_port=8052"' -- run -m co.paralleluniverse.pulsar.examples.cluster.ping

Makes the console print something like "Waiting for pong to be registered"

Then, after running the pong startup command I get:

Pong started

21:30:47.763 [udpCommReceiveExecutor-6]                  netty.UDPComm [WARN ] {} Exception caught in channel [id: 0x36d1504c, 0.0.0.0/0.0.0.0:7050]:  java.lang.RuntimeException Message received from unhandled node -1 java.lang.RuntimeException: Message received from unhandled node -1

at co.paralleluniverse.galaxy.netty.UDPComm.messageReceived(UDPComm.java:454)

at co.paralleluniverse.galaxy.netty.UDPComm$1$1.messageReceived(UDPComm.java:299)

at co.paralleluniverse.galaxy.netty.ChannelMessageNodeResolver.messageReceived(ChannelMessageNodeResolver.java:36)

at co.paralleluniverse.galaxy.netty.OneToOneCodec.handleUpstream(OneToOneCodec.java:63)

at org.jboss.netty.handler.execution.ChannelUpstreamEventRunnable.doRun(ChannelUpstreamEventRunnable.java:43)

at org.jboss.netty.handler.execution.ChannelEventRunnable.run(ChannelEventRunnable.java:67)

at org.jboss.netty.handler.execution.OrderedMemoryAwareThreadPoolExecutor$ChildExecutor.run(OrderedMemoryAwareThreadPoolExecutor.java:314)

21:30:47.817 [udpCommReceiveExecutor-1]                  netty.UDPComm [WARN ] {} Exception caught in channel [id: 0x36d1504c, 0.0.0.0/0.0.0.0:7050]:  java.lang.RuntimeException Message received from unhandled node -1 java.lang.RuntimeException: Message received from unhandled node -1

at co.paralleluniverse.galaxy.netty.UDPComm.messageReceived(UDPComm.java:454)

at co.paralleluniverse.galaxy.netty.UDPComm$1$1.messageReceived(UDPComm.java:299)

at co.paralleluniverse.galaxy.netty.ChannelMessageNodeResolver.messageReceived(ChannelMessageNodeResolver.java:36)

at co.paralleluniverse.galaxy.netty.OneToOneCodec.handleUpstream(OneToOneCodec.java:63)

at org.jboss.netty.handler.execution.ChannelUpstreamEventRunnable.doRun(ChannelUpstreamEventRunnable.java:43)

at org.jboss.netty.handler.execution.ChannelEventRunnable.run(ChannelEventRunnable.java:67)

at org.jboss.netty.handler.execution.OrderedMemoryAwareThreadPoolExecutor$ChildExecutor.run(OrderedMemoryAwareThreadPoolExecutor.java:314)

21:30:47.932 [pool-9-thread-1]                     core.Cache [WARN ] {} TIMEOUT: Op.GET(line:3497f8, data:co.paralleluniverse.galaxy.core.StringRootManager$StringRootPageHandler@716cb856) 

21:30:47.934 [ForkJoinPool-default-fiber-pool-worker-5]       galaxy.GlxGlobalRegistry [ERROR] {} Registering actor LocalActorRef{actor: PulsarActor@pong[owner: fiber-10000001]} at root pong failed due to timeout 

Exception in thread "main" java.lang.RuntimeException: Actor registration failed

at co.paralleluniverse.remote.galaxy.GlxGlobalRegistry.register(GlxGlobalRegistry.java:74)

There is no changes. It is a simples checkout/run procedure. And no expected outcome.

Running the pingpong-register with a single instance, all the thing works as expected. 

Adding a galaxy server to scenari, I get the following output on node startups:

21:36:46.364 [jgroups-0]           core.AbstractCluster [WARN ] {} No reader set for property ip_server_port (found in node freeplastic-28313) 

21:36:46.366 [jgroups-0]           core.AbstractCluster [WARN ] {} No reader set for property ip_server_port (found in node freeplastic-28313) 

21:36:46.367 [jgroups-0] netty.SocketNodeAddressResolver [WARN ] {} Socket port (property ip_port) not set for node NODE freeplastic-28313 id: 0 ip_addr: /127.0.1.1 jg_addr: freeplastic-28313 

I don't know. Tried to combine two different machines, one for server other for peers. Nothing happens. If anybody managed to run the very first examples, please share what you did.

regards

--
Paulo Suzart
@paulosuzart

[pron]

Hi Paulo.

It seems like jgroups can't function properly. Is multicast enabled on your cluster? If not, it's best to use ZooKeeper instead of jgroups; in that case, let me know and I'll guide you through it.

Ron

[Paulo Suzart]

Thanks Ron, Yes my network/machines are supposed to work with multicast. But switching to ZK:

After starting a server and two nodes zookeeper shows:

[zk: localhost:2181(CONNECTED) 8] ls /co.paralleluniverse.galaxy/nodes                

[node-0000000000, node-0000000002, node-0000000001]

But the nodes spit: 21:45:44.753 [main-EventThread]           zookeeper.ClientCnxn [ERROR] {} Error while calling watcher  java.lang.AssertionError: java.io.StreamCorruptedException: invalid stream header: 3132372E

at co.paralleluniverse.galaxy.zookeeper.ZooKeeperDistributedTree$EphemeralChildren.fromByteArray(ZooKeeperDistributedTree.java

Same for server. I tried a docker container with zookeeper and got the same result.

This StreamCorruptedException is very strange. I'm using [org.apache.curator/curator-framework "2.4.2"]

Any idea?

Cheers

--
You received this message because you are subscribed to the Google Groups "quasar-pulsar-user" group.
To unsubscribe from this group and stop receiving emails from it, send an email to quasar-pulsar-u...@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

--
Paulo Suzart
@paulosuzart

[pron]

In the upcoming version of Quasar (0.6.0) we're concentrating on distributed configurations on top of Galaxy. I don't yet know what causes your problem (though it does appear to be some network configuration issue), but we'll probably hit on that as work on 0.6.0 progresses.

Ron

On Saturday, May 17, 2014 3:51:24 AM UTC+3, Paulo Suzart wrote:

Thanks Ron, Yes my network/machines are supposed to work with multicast. But switching to ZK:

After starting a server and two nodes zookeeper shows:

[zk: localhost:2181(CONNECTED) 8] ls /co.paralleluniverse.galaxy/nodes                

[node-0000000000, node-0000000002, node-0000000001]

But the nodes spit: 21:45:44.753 [main-EventThread]           zookeeper.ClientCnxn [ERROR] {} Error while calling watcher  java.lang.AssertionError: java.io.StreamCorruptedException: invalid stream header: 3132372E

at co.paralleluniverse.galaxy.zookeeper.ZooKeeperDistributedTree$EphemeralChildren.fromByteArray(ZooKeeperDistributedTree.java

Same for server. I tried a docker container with zookeeper and got the same result.

This StreamCorruptedException is very strange. I'm using [org.apache.curator/curator-framework "2.4.2"]

Any idea?

Cheers

To unsubscribe from this group and stop receiving emails from it, send an email to quasar-pulsar-user+unsub...@googlegroups.com.

For more options, visit https://groups.google.com/d/optout.

--
Paulo Suzart
@paulosuzart

[pron]

Sure enough, we've stumbled on this bug while working on Quasar 0.6.0 (which, along with a new Galaxy version) will be released in a week or two. The problem is a recent Curator version that's changed the default value for new ZK nodes (from all zeros to some non-zero value). 

[Tao]

Hi Ron,

I just tried the latest pingpong sample from https://github.com/puniverse/pulsar/tree/master/src/test/clojure/co/paralleluniverse/pulsar/examples/cluster and still got the same error. And I have tried on both Mac OSX 10.10.5 and Ubuntu 15.10 and got the same 'Message received from unhandled node -1' error. I don't want to use ZooKeeper because we don't want introduce any dependence. 

One more question: Is Galaxy ready for production or still in development? Because cluster actor is key for us.

Regards,

Tao