This is a rough one. We have a client who had a NetD 4.1
deployment server on a certain machine. As their user
base increased, the machine was deemed insufficient and a
different machine was set up at the same IP address.
Unfortunately, the guy who installed the server on the
new machine changed the name of the default partition.
When a user who bookmarked an old link hits the server,
the "peServiceId" value within the SPIDERSESSION web value
is incorrect. This causes the CNdPEProxyImpl to catch a
CNdNoServiceException generated by a call to the getService()
method of the parent CNdServiceProxy. No problem, right?
The server logs an error and displays a software message
indicating that the specified service could not be located.
The problem is that every subsequent access to the CP that
handled the error results in the same situation. The
old link seems to cause the CP to "lose" the PE. I cannot
trap the bad peServiceId, because the damage occurs before
any project level events are triggered. The session is null
before my code is even referenced. A new session cannot be
created, as the PE proxies are already invalid. The only
way to rectify the problem is to restart the service
manager. We currently have our client monitoring the server
around the clock, and immediately restarting the SM when
the PE proxy refs evaporate. Note that the PE is alive and
well during this whole disaster. This is a major crisis for
all involved. Any thoughts on the subject would be greatly
appreciated.
-Sam Koppes
_________________________________________________________________________
For help in using, subscribing, and unsubscribing to the discussion
forums, please go to: http://www.netdynamics.com/support/visitdevfor.html
For dire need help, email: [EMAIL PROTECTED]
