random timeout with digital sign, DocumentManagerClient.java:329

[#1]

hey guys,

Not sure if any one of you have similar experience. We are running Livecyle.ES 8.2.1 and so far we use this for signing PDF files only for a government agency.We use EJB endpoint.

Our application is a custom webapp which use Livecycle Java AIP. The problem we see is on and off, we see this exception during sign PDF.

2009-04-04 10:37:50,420 ERROR [STDERR] Apr 4, 2009 10:37:50 AM com.adobe.idp.DocumentManagerClient requestRemotePassivation
SEVERE: DOCS001: Unexpected exception. See the stack trace for details.
org.omg.CORBA.COMM_FAILURE:   vmcid: SUN  minor code: 208 completed: Maybe
    at com.sun.corba.se.impl.logging.ORBUtilSystemException.connectionAbort(ORBUtilSystemException.java:2372)

And this error is raised from,

com.adobe.idp.DocumentError: org.omg.CORBA.COMM_FAILURE:   vmcid: SUN  minor code: 208 completed: Maybe
    at com.adobe.idp.DocumentManagerClient.requestRemotePassivation(DocumentManagerClient.java:329)
    at com.adobe.idp.Document.passivate(Document.java:1069)
    at com.adobe.idp.Document.passivate(Document.java:1012)
    at com.adobe.idp.Document.getInputStream(Document.java:1524)

The interesting this about this error is, documents get signed before and after this error, are just fine. This problem used to last about 1 minute and gone. And this happens only in our PROD env . I dont see this in my DEV or TEST. The tricky thing is, in PROD, we connect to a hardware signature profile, rather than a normal scenario, not sure if related.

My guessing is that the document do get signed on the server side, and then somehow when client request it, it is gone or something.

Well, i am stomped. Really appreciate if anyone of you can share some insight.

Thanks.

-ying

[#2]

Well, after a week's test I have a walkaround for this problem, I changed code to use SOAP endpoint instead of EJB. I have to do this for another reason.

My application is a webservice running under another JBoss instance. Before, I have to restart JBoss, each time, if Livecycle get bounced. This, I guess has to do with ServiceFactory cached javax.naming.InitialContext. Once Livecycle get bounced, the InitialContext get obsolete. The good part of using a SOAP endpoint is, I dont need bounce my JBoss any more.

After I made the switch, this old random timeout problem seems like going away. I still not fully understand the reason. One thing is a fact that for EJB, I have to get permission to open 1099 port through firewall (two of them). But, for SOAP, 8080 also need be opened. I am not a network kind of person, maybe HTTP is treated differently?

Well, well, as long as I dont see the problem, I let it go ...

-ying

[tareknas]

I need your help. In my case, Java code is invoking the API myData.passivate() which fails with the same error. This happens only after the Java Portal is left for a couple of hour not used. If I restart Tomcat server or try again after the error, it works fine. Please hlep.