Why does manager print an error stack trace that begins with "Failed to invoke 'components_unavailable()"?

Sometimes manager prints out enormous error stack traces complaining that it cannot invoke some method of a client such as this:

[ Connected to 146.88.1.153:57431 ]
2004-08-03T21:57:01.300 Client 'ControlTestR0a' logged in.

#...

[ Retrying to connect to 146.88.1.153:57431 ]
[ Retrying to connect to 146.88.1.153:57431 ]
[ Retrying to connect to 146.88.1.153:57431 ]
[ Retrying to connect to 146.88.1.153:57431 ]
[ Retrying to connect to 146.88.1.153:57431 ]
2004-08-03T21:57:46.882 RemoteException caught while invoking 'Client.components_unavailable' on ClientInfo =3D { name =3D 'ControlTestR0a', Handle (0x2d60001) =3D { type =3D CLIENT, key =3D (0xd6), id =3D (0x1) } }.
[ Retrying to connect to 146.88.1.153:57431 ]
[ Retrying to connect to 146.88.1.153:57431 ]
[ Retrying to connect to 146.88.1.153:57431 ]
[ Retrying to connect to 146.88.1.153:57431 ]
2004-08-03T21:57:46.876 Failed to invoke 'components_unavailable()' method.
2004-08-03T21:57:48.887 Exception report:
-------------------->( 0 )<--------------------
        org.omg.CORBA.TRANSIENT: Retries exceeded, couldn't reconnect to 146.88.1.153:57431  vmcid: 0x0  minor code: 0  completed: No
                at org.jacorb.orb.connection.Client_TCP_IP_Transport.connect(Unknown Source)
                at org.jacorb.orb.connection.TCP_IP_Transport.write(Unknown Source)
                at org.jacorb.orb.connection.GIOPConnection.write(Unknown Source)
                at org.jacorb.orb.CDROutputStream.write(Unknown Source)
                at org.jacorb.orb.connection.ServiceContextTransportingOutputStream.write_to(Unknown Source)
                at org.jacorb.orb.connection.GIOPConnection.sendMessage(Unknown Source)
                at org.jacorb.orb.connection.ClientConnection.sendRequest(Unknown Source)
                at org.jacorb.orb.Delegate.invoke(Unknown Source)
                at org.omg.CORBA.portable.ObjectImpl._invoke(ObjectImpl.java:457)
                at si.ijs.maci._ClientStub.components_unavailable(_ClientStub.java:31)
                at com.cosylab.acs.maci.plug.ClientProxy.components_unavailable(ClientProxy.java:131)
                at com.cosylab.acs.maci.manager.ManagerImpl$1ComponentUnavailableTask.run(ManagerImpl.java:4050)
                at EDU.oswego.cs.dl.util.concurrent.PooledExecutor$Worker.run(PooledExecutor.java:728)
                at java.lang.Thread.run(Thread.java:534)
--------------------------------------------------

#...

Answer:

A client that has connected to manager did not disconnect properly following the guidlines ACS has set fourth in documentation and examples. In simple terms, either:

• the disconnect method of a Python PySimpleClient object was not invoked before exiting the Python interpreter or
• the cleanUp method of a Java alma.acs.component.client.ComponentClient object was not invoked before exiting the JVM

-- DavidFugate - 14 Aug 2004