Olivier Pernet wrote: > PS. I had done something wrong with the hostname, now I'm back to > seeing the connection to the EC2 instance, but having the deployment > hang and eventually timeout. > > Using the RMI debug properties sun.rmi.transport.logLevel=VERBOSE and > java.rmi.server.logCalls=true, this is all I can see on the server: > > SmartFrog ready... [rootProcess:3800] Tue Sep 16 19:08:27 UTC 2008 > Sep 16, 2008 7:11:10 PM sun.rmi.transport.StreamRemoteCall getInputStream > FINER: RMI TCP Connection(1)-193.1.133.1: getting input stream > Sep 16, 2008 7:11:10 PM sun.rmi.transport.Transport serviceCall > FINER: RMI TCP Connection(1)-193.1.133.1: call dispatcher > Sep 16, 2008 7:11:10 PM sun.rmi.server.UnicastServerRef logCall > FINER: RMI TCP Connection(1)-193.1.133.1: [193.1.133.1: > sun.rmi.registry.RegistryImpl[0:0:0, 0]: java.rmi.Remote > lookup(java.lang.String)] > Sep 16, 2008 7:11:10 PM sun.rmi.transport.StreamRemoteCall getOutputStream > FINER: RMI TCP Connection(1)-193.1.133.1: getting output stream >
We may be up against the whole Java RMI infrastructure here. I wonder if the runtime on the machine, once it is live and listening for connections, isn't able to talk to itself any more because the IP address is wrong. Nat is getting in the way. 1. Have you tried setting up an SSH tunnel instead. So you, at your desk, set up a port to tunnel over to the EC2 host, then deploy to your local host. 2. The long haul deployment I've done has only used the WS-* based CDDLM server and client, which hasn't been used for a while and may have some surprises, even if you do want to use it for long-haul work. In particular, getting new JARs up on the remote classpath can be tricky. 3. The Rest component does have a basic REST API, though I dont think it is complete. Doing a long haul "High" REST API is something I've long wanted to do, just been distracted by so many other things. Interestingly enough, the problem also arises in Hadoop once you add a lifecycle to its services https://issues.apache.org/jira/browse/HADOOP-3969 That bugrep shows what I want in a REST API to SmartFrog - manipulate the state of a component by putting it into a specific state - have all attributes exported/manipulable - links to other components become URLs The component graph effectively becomes the hypermedia representation of the system state. There's one big catch here: the existing RMI-style methods. > > And on the daemon running on my PC, with sun.rmi.client.logCalls=true, > sun.rmi.server.logLevel=VERBOSE, sun.rmi.transport.logLevel=VERBOSE > (early messages omitted): > > 16-Sep-2008 20:11:10 sun.rmi.server.UnicastRef logClientCall > FINER: RMI TCP Connection(3)-193.1.133.1: outbound call: > [endpoint:[75.101.228.74:3800](remote),objID:[0:0:0, 0]] : > sun.rmi.registry.RegistryImpl_Stub[0:0:0, 0]: java.rmi.Remote > lookup(java.lang.String) > 16-Sep-2008 20:11:10 sun.rmi.transport.StreamRemoteCall <init> > FINER: RMI TCP Connection(3)-193.1.133.1: write remote call header... > 16-Sep-2008 20:11:10 sun.rmi.transport.StreamRemoteCall getOutputStream > FINER: RMI TCP Connection(3)-193.1.133.1: getting output stream > 16-Sep-2008 20:11:10 sun.rmi.server.UnicastRef invoke > FINER: RMI TCP Connection(3)-193.1.133.1: execute call > 16-Sep-2008 20:11:10 sun.rmi.transport.StreamRemoteCall getInputStream > FINER: RMI TCP Connection(3)-193.1.133.1: getting input stream > 16-Sep-2008 20:11:10 sun.rmi.server.UnicastRef done > FINE: RMI TCP Connection(3)-193.1.133.1: free connection (reuse = true) > 16-Sep-2008 20:11:10 sun.rmi.server.UnicastRef newCall > FINE: RMI TCP Connection(3)-193.1.133.1: get connection > 16-Sep-2008 20:14:19 sun.rmi.server.UnicastRef newCall > FINE: RMI RenewClean-[ec2-75-101-228-74.compute-1.amazonaws.com:48712]: > get connection > 16-Sep-2008 20:14:19 sun.rmi.server.UnicastRef invoke > FINER: RMI TCP Connection(3)-193.1.133.1: method: public abstract > org.smartfrog.sfcore.prim.Prim > org.smartfrog.sfcore.compound.Compound.sfDeployComponentDescription(java.lang.Object,org.smartfrog.sfcore.prim.Prim,org.smartfrog.sfcore.componentdescription.ComponentDescription,org.smartfrog.sfcore.common.Context) > throws > java.rmi.RemoteException,org.smartfrog.sfcore.common.SmartFrogDeploymentException > 16-Sep-2008 20:14:19 sun.rmi.server.UnicastRef logClientCall > FINER: RMI TCP Connection(3)-193.1.133.1: outbound call: > [endpoint:[ec2-75-101-228-74.compute-1.amazonaws.com:48712](remote),objID:[d5a148a:11c6c9045b8:-7fff, > 324988096577243388]] : > org.smartfrog.sfcore.processcompound.ProcessCompoundImpl_Stub[d5a148a:11c6c9045b8:-7fff, > 324988096577243388]: public abstract org.smartfrog.sfcore.prim.Prim > org.smartfrog.sfcore.compound.Compound.sfDeployComponentDescription(java.lang.Object,org.smartfrog.sfcore.prim.Prim,org.smartfrog.sfcore.componentdescription.ComponentDescription,org.smartfrog.sfcore.common.Context) > throws > java.rmi.RemoteException,org.smartfrog.sfcore.common.SmartFrogDeploymentException > 16-Sep-2008 20:16:10 sun.rmi.transport.StreamRemoteCall getInputStream > FINER: RMI TCP Connection(4)-193.1.133.1: getting input stream > 16-Sep-2008 20:16:10 sun.rmi.transport.Transport serviceCall > FINER: RMI TCP Connection(4)-193.1.133.1: call dispatcher > 16-Sep-2008 20:16:10 sun.rmi.transport.StreamRemoteCall getOutputStream > FINER: RMI TCP Connection(4)-193.1.133.1: getting output stream > 2008/09/16 20:17:29:087 IST [ERROR][RMI TCP Connection(3)-193.1.133.1] > SFCORE_LOG - caught on start (HOST > tmonahan.ucd.ie:rootProcess:sim:servers) > <SmartFrogLifecycleException:: [sfStart] HOST > tmonahan.ucd.ie:rootProcess:sim:servers:workflow > cause: SmartFrogDeploymentException: unnamed component. > SmartFrogLifecycleException:: [sfStart] Failed to start a new child > 'prepare'., cause: SmartFrogDeploymentException: 'pwdServers' > component. PrimHostDeployerImpl.deploy, cause: > java.rmi.ConnectException: Connection refused to host: > ec2-75-101-228-74.compute-1.amazonaws.com; nested exception is: > java.net.ConnectException: Connection timed out, SmartFrog > 3.12.041dev (2008-08-13 18:58:16 IST), primSFCompleteName: HOST > tmonahan.ucd.ie:rootProcess:sim:servers:workflow, primContext: > included, reference: HOST > tmonahan.ucd.ie:rootProcess:sim:servers:workflow, primContext: > included > cause: SmartFrogLifecycleException:: [sfStart] Failed to start > a new child 'prepare'. > cause: SmartFrogDeploymentException: 'pwdServers' component. > PrimHostDeployerImpl.deploy > cause: java.rmi.ConnectException: Connection refused to > host: ec2-75-101-228-74.compute-1.amazonaws.com; nested exception is: > java.net.ConnectException: Connection timed out > ------------------------------------------------------------------------- This SF.Net email is sponsored by the Moblin Your Move Developer's challenge Build the coolest Linux based applications with Moblin SDK & win great prizes Grand prize is a trip for two to an Open Source event anywhere in the world http://moblin-contest.org/redirect.php?banner_id=100&url=/ _______________________________________________ Smartfrog-users mailing list [email protected] https://lists.sourceforge.net/lists/listinfo/smartfrog-users
