Hello,
I'm looking to move our distributed services (currently using Camel's JMS
remoting) to a dOSGi solution. I won't be looking at ECF (simply because
they don't seem to publish up-to-date maven artifacts), so that leaves Fuse
Fabric and Karaf Cellar.
Apologies for the length of this mail, what follows is a log of the steps I
took with Fabric, in the hope that somebody can spot something.
Thought I'd try to get something working in FuseESB before porting in back
to Karaf 2.2.x.
So far I've burnt a bit of time with Fabric but not made any progress
(tried trunk and fuse-fabric-7.0.0.fuse-061, currently looking at
fuse-fabric-7.0.1.fuse-084). I suspect I'm doing something stupid or have
missed something vital out.
Starting with the simplest distributed case I've got a features.xml and 3
simple bundles; a service interface bundle, a client bundle and a service
implementation bundle. I've ensured the client and service work in a
single instance, but am a bit stumped as to how to get this running over
fabric.
The service imp exports the property "service.exported.interfaces" as "*"
and the interface implied by the asterisk is available to both service imp
and client bundles (i.e. deployed to both instances).
On both boxes I've temporarily instructed iptables to open all ports and
I'm running the fusefabric command as root.
# On box1 (provider):
tar -xzf /data/share/fuse-fabric-7.0.1.fuse-084.tar.gz
cd fuse-fabric-7.0.1.fuse-084/
sudo ./bin/fusefabric
# On box1's Karaf shell:
log:set TRACE net.earcam
log:set TRACE org.fusesource.fabric.dosgi
features:install fabric-commands fabric-dosgi fabric-configadmin
fabric-zookeeper-commands
# Line above logs exception as Zookeeper not created or ensembled -
assuming safe to ignore
fabric:create --clean root # Causes exception 1 (see below), again
assuming this is safe to ignore
fabric:profile-create --parents dosgi dosgi-provider
fabric:profile-edit --repositories
mvn:net.earcam.example.hello/net.earcam.example.hello.feature/0.0.1-SNAPSHOT/xml/features
dosgi-provider
fabric:profile-edit --features hello-server dosgi-provider
fabric:container-create --profile dosgi-provider --parent root
dosgi-provider
fabric:container-list # Shows success
fabric:container-connect dosgi-provider # Connects fine, can see
..example.hello bundles installed
log:set TRACE net.earcam
log:set TRACE org.fusesource.fabric.dosgi
# ------
# On box2 (consumer):
tar -xzf /data/share/fuse-fabric-7.0.1.fuse-084.tar.gz
cd fuse-fabric-7.0.1.fuse-084/
sed -i s/karaf.name=root/karaf.name=vm-mint1/ etc/system.properties
sudo ./bin/fusefabric
# On box2's Karaf shell:
log:set TRACE net.earcam
log:set TRACE org.fusesource.fabric.dosgi
features:install fabric-commands fabric-dosgi fabric-configadmin
fabric-zookeeper-commands
fabric:join 10.39.216.49 # The IP of box1
fabric:profile-create --parents dosgi dosgi-consumer
fabric:profile-edit --repositories
mvn:net.earcam.example.hello/net.earcam.example.hello.feature/0.0.1-SNAPSHOT/xml/features
dosgi-consumer
fabric:profile-edit --features hello-client dosgi-consumer
fabric:container-create --profile dosgi-consumer --parent vm-mint1
dosgi-consumer
fabric:container-list # Shows success
fabric:container-connect dosgi-consumer # Connects fine, can see
..example.hello bundles installed
log:set TRACE net.earcam
log:set TRACE org.fusesource.fabric.dosgi
# ------
The "provision status" is "success" for all when running container-list and
I can connect to the two new containers dosgi-provider and dosgi-consumer.
On box2, in the dosgi-consumer container, I can see the remote service
exported from box1, and it disappears/appears as I stop/start the exporting
service bundle on box1.
The imported service has "correct looking" properties:
component.id = 2
component.name = net.earcam.example.hello.server.HelloService
endpoint.framework.uuid = b5806ad9-7849-4aaf-adf7-56ba1e6538e3
endpoint.id = 2008920412-47095-1348493938958-0-3
fabric.address = tcp://localhost:44462
objectClass = net.earcam.example.hello.api.GreetingService
service.id = 246
service.imported = true
service.imported.configs = fabric-dosgi
service.pid = net.earcam.example.hello.server.HelloService
# Exception 1, seen on both boxes:
java.lang.NullPointerException
at org.fusesource.fabric.dosgi.impl.Manager.destroy(Manager.java:175)
at org.fusesource.fabric.dosgi.Activator.destroy(Activator.java:46)
at sun.reflect.NativeMethodAccessorImpl.invoke0(Native
Method)[:1.6.0_33]
at
sun.reflect.NativeMethodAccessorImpl.invoke(NativeMethodAccessorImpl.java:39)[:1.6.0_33]
at
sun.reflect.DelegatingMethodAccessorImpl.invoke(DelegatingMethodAccessorImpl.java:25)[:1.6.0_33]
at java.lang.reflect.Method.invoke(Method.java:597)[:1.6.0_33]
at
org.apache.aries.blueprint.utils.ReflectionUtils.invoke(ReflectionUtils.java:225)[10:org.apache.aries.blueprint:0.3.1.fuse-70-084]
at
org.apache.aries.blueprint.container.BeanRecipe.invoke(BeanRecipe.java:838)[10:org.apache.aries.blueprint:0.3.1.fuse-70-084]
at
org.apache.aries.blueprint.container.BeanRecipe.destroy(BeanRecipe.java:743)[10:org.apache.aries.blueprint:0.3.1.fuse-70-084]
at
org.apache.aries.blueprint.container.BlueprintRepository.destroy(BlueprintRepository.java:295)[10:org.apache.aries.blueprint:0.3.1.fuse-70-084]
at
org.apache.aries.blueprint.container.BlueprintContainerImpl.destroyComponents(BlueprintContainerImpl.java:673)[10:org.apache.aries.blueprint:0.3.1.fuse-70-084]
at
org.apache.aries.blueprint.container.BlueprintContainerImpl.destroy(BlueprintContainerImpl.java:826)[10:org.apache.aries.blueprint:0.3.1.fuse-70-084]
at
org.apache.aries.blueprint.container.BlueprintExtender.destroyContext(BlueprintExtender.java:255)[10:org.apache.aries.blueprint:0.3.1.fuse-70-084]
at
org.apache.aries.blueprint.container.BlueprintExtender.bundleChanged(BlueprintExtender.java:247)[10:org.apache.aries.blueprint:0.3.1.fuse-70-084]
at
org.apache.aries.blueprint.container.BlueprintExtender$BlueprintBundleTrackerCustomizer.modifiedBundle(BlueprintExtender.java:471)[10:org.apache.aries.blueprint:0.3.1.fuse-70-084]
at
org.osgi.util.tracker.BundleTracker$Tracked.customizerModified(BundleTracker.java:495)[karaf.jar:2.2.5.fuse-70-084]
at
org.osgi.util.tracker.BundleTracker$Tracked.customizerModified(BundleTracker.java:1)[karaf.jar:2.2.5.fuse-70-084]
at
org.osgi.util.tracker.AbstractTracked.track(AbstractTracked.java:238)[karaf.jar:2.2.5.fuse-70-084]
at
org.osgi.util.tracker.BundleTracker$Tracked.bundleChanged(BundleTracker.java:457)[karaf.jar:2.2.5.fuse-70-084]
at
org.apache.felix.framework.util.EventDispatcher.invokeBundleListenerCallback(EventDispatcher.java:870)[org.apache.felix.framework-4.0.3.fuse-70-084.jar:]
at
org.apache.felix.framework.util.EventDispatcher.fireEventImmediately(EventDispatcher.java:791)[org.apache.felix.framework-4.0.3.fuse-70-084.jar:]
at
org.apache.felix.framework.util.EventDispatcher.fireBundleEvent(EventDispatcher.java:515)[org.apache.felix.framework-4.0.3.fuse-70-084.jar:]
at
org.apache.felix.framework.Felix.fireBundleEvent(Felix.java:4431)[org.apache.felix.framework-4.0.3.fuse-70-084.jar:]
at
org.apache.felix.framework.Felix.stopBundle(Felix.java:2532)[org.apache.felix.framework-4.0.3.fuse-70-084.jar:]
at
org.apache.felix.framework.BundleImpl.stop(BundleImpl.java:983)[org.apache.felix.framework-4.0.3.fuse-70-084.jar:]
at
org.fusesource.fabric.agent.DeploymentAgent.updateDeployment(DeploymentAgent.java:709)[102:org.fusesource.fabric.fabric-agent:7.0.1.fuse-084]
at
org.fusesource.fabric.agent.DeploymentAgent.doUpdate(DeploymentAgent.java:415)[102:org.fusesource.fabric.fabric-agent:7.0.1.fuse-084]
at
org.fusesource.fabric.agent.DeploymentAgent$1.run(DeploymentAgent.java:225)[102:org.fusesource.fabric.fabric-agent:7.0.1.fuse-084]
at
java.util.concurrent.Executors$RunnableAdapter.call(Executors.java:441)[:1.6.0_33]
at
java.util.concurrent.FutureTask$Sync.innerRun(FutureTask.java:303)[:1.6.0_33]
at java.util.concurrent.FutureTask.run(FutureTask.java:138)[:1.6.0_33]
at
java.util.concurrent.ThreadPoolExecutor$Worker.runTask(ThreadPoolExecutor.java:886)[:1.6.0_33]
at
java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:908)[:1.6.0_33]
at java.lang.Thread.run(Thread.java:662)[:1.6.0_33]
# Exception 2, seen on box2:
2012-09-24 15:16:00,142 | INFO | spatch-DEFAULT-1 |
TransportPool | 67 - org.fusesource.fabric.fabric-dosgi
- 7.0.1.fuse-084 | Transport failure
java.net.ConnectException: Connection refused
at sun.nio.ch.SocketChannelImpl.checkConnect(Native
Method)[:1.7.0_07]
at
sun.nio.ch.SocketChannelImpl.finishConnect(SocketChannelImpl.java:692)[:1.7.0_07]
at
org.fusesource.fabric.dosgi.tcp.TcpTransport$5.run(TcpTransport.java:483)[67:org.fusesource.fabric.fabric-dosgi:7.0.1.fuse-084]
at
org.fusesource.hawtdispatch.internal.NioDispatchSource$3.run(NioDispatchSource.java:228)[78:org.fusesource.hawtdispatch.hawtdispatch:1.9.0]
at
org.fusesource.hawtdispatch.internal.SerialDispatchQueue.run(SerialDispatchQueue.java:84)[78:org.fusesource.hawtdispatch.hawtdispatch:1.9.0]
at
org.fusesource.hawtdispatch.internal.pool.SimpleThread.run(SimpleThread.java:77)[78:org.fusesource.hawtdispatch.hawtdispatch:1.9.0]
2012-09-24 15:16:00,147 | INFO | agent-1-thread-1 |
DeploymentAgent | 63 - org.fusesource.fabric.fabric-agent
- 7.0.1.fuse-084 | Done.
# After which the logs are filled with:
2012-09-24 15:18:14,435 | INFO | nana.local:2181) |
ClientCnxn | 54 -
org.fusesource.fabric.fabric-linkedin-zookeeper - 7.0.1.fuse-084 | Opening
socket connection to server banana.local/10.39.216.49:2181
2012-09-24 15:18:14,436 | INFO | nana.local:2181) |
ZooKeeperSaslClient | 54 -
org.fusesource.fabric.fabric-linkedin-zookeeper - 7.0.1.fuse-084 | Client
will not SASL-authenticate because the default JAAS configuration section
'Client' could not be found. If you are not using SASL, you may ignore
this. On the other hand, if you expected SASL to work, please fix your JAAS
configuration.
2012-09-24 15:18:14,437 | INFO | nana.local:2181) |
ClientCnxn | 54 -
org.fusesource.fabric.fabric-linkedin-zookeeper - 7.0.1.fuse-084 | Socket
connection established to banana.local/10.39.216.49:2181, initiating session
2012-09-24 15:18:14,438 | INFO | nana.local:2181) |
ClientCnxn | 54 -
org.fusesource.fabric.fabric-linkedin-zookeeper - 7.0.1.fuse-084 | Unable
to read additional data from server sessionid 0x139f880728d0005, likely
server has closed socket, closing socket connection and attempting reconnect
>From what I've read the SASL messages are safe to ignore, but exception 2
looks related to the problem - I don't think it's a configuration issue
with the OSes, the SocketChannel fails to connect but all ports are open
and I can telnet remotely to that ip on port 2181.
The code I'm using can be found here
https://dl.dropbox.com/u/2465717/net.earcam.example.hello.tgz (just a
simple maven project).
I'd be very grateful if anyone has a suggestions as to the cause, an RTFM
with link, or a pointer to a working example, etc.
thanks,
Best regards,
Caspar
