Name: krC82822 Date: 02/10/2001


10 Feb 2001, eval1127@eng -- this may really be a doc bug/RFE
(i.e., we may need to make removal of the previous log more explicit in our docs).
See also # 4320415.
--------------------
java version "1.2.2"
Classic VM (build JDK-1.2.2-001, green threads, sunwjit)
[and 1.3, based on bugs such as # 4320415]

During development, [an] RMI-based application server is occsionally restarted.
After changing the implementation of the setup class that registers with
rmid, the class would no longer be able to create its application group:
===========================================================================
Exception in thread "main" java.rmi.activation.ActivationException: ActivationSy
stem not running; nested exception is:
        java.rmi.NotBoundException: java.rmi.activation.ActivationSystem
java.rmi.NotBoundException: java.rmi.activation.ActivationSystem
        at sun.rmi.transport.StreamRemoteCall.exceptionReceivedFromServer(Stream
RemoteCall.java:249)
        at sun.rmi.transport.StreamRemoteCall.executeCall(StreamRemoteCall.java:224)
        at sun.rmi.server.UnicastRef.invoke(UnicastRef.java:358)
        at sun.rmi.registry.RegistryImpl_Stub.lookup(Unknown Source)
        at java.rmi.Naming.lookup(Naming.java:89)
        at java.rmi.activation.ActivationGroup.getSystem(ActivationGroup.java:441)
        at avalon.dataaccess.remotedatasource.AASSetup.main(AASSetup.java:60)
================================================================================

This problem disappears when the old ./log directory is deleted before
rmiregistry and rmid are restarted. It does depend on the contents of the
file 'Logfile.1' which seems to contain serialized objects.

rmid does NOT produce any error message when this happens.

---------------

23 Jan 2001, eval1127@eng -- user notes in follow-up email:

"Actually I did review the bug database BEFORE submitting the bug and
considered it a new one. The reason is that
I actually traced it to the log files and there is a workaround that fixes
the problem for me.
The workaround is the reason for classifying it as 'not preventing me from
continuing my work'.

However, the entry into the bug database would help others who might run
inot the same problem and save them
a lot of hassle."
(Review ID: 115605)
======================================================================