Problem using Async20Access in Utgard: 'Oxid Resolver Thread: Socket Closed'

[Daniel B.]

Hello,

I'm using Async20Access group of Utgard 1.4 (Github) driver in my app. After I create the group updates start to receive normally, but every 2 minutes I see j-interop warnings (even after stop receiving updates, see below):

19:22:43,335 WARN  org.jinterop              Mar 31, 2016 7:22:43 PM org.jinterop.dcom.core.JIComOxidRuntimeHelper$1 run
WARNING: Oxid Resolver Thread: Socket Closed , on thread Id: jI_OxidResolver_Client[51977 , 49388]
     -     -     -     -     -     -     -     -     -     -     [jI_OxidResolver_Client[51977 , 49388]]  at java.util.logging.Logger.log(Logger.java:616)
 at java.util.logging.Logger.doLog(Logger.java:641)
 at java.util.logging.Logger.log(Logger.java:664)
 at java.util.logging.Logger.warning(Logger.java:1166)
 at org.jinterop.dcom.core.JIComOxidRuntimeHelper$1.run(JIComOxidRuntimeHelper.java:93)
 at java.lang.Thread.run(Thread.java:745)

AFAIU it is DCOM ping thread that is killed by (probably) async group, because there are no warnings when polling an OPC server in sync mode.

But the most sad thing is that updates stop to receive after some time, 15-20 minutes usually. The only advice I found is to turn off CoClass auto collection in j-interop, like this:

JISystem.setJavaCoClassAutoCollection(false);

But this also does not help, besides that async updates lost period is greater 8mins as described in advice.

The OPC server we're connecting to works in async mode only, but can be reproduced with Archestra FSGateway for example.

So the problem is really painful, it would be great if someone could give us a clue about that.

[Jens Reimann]

The only clue I can give you is that we never used the async mode in production, because it simply is too unstable.

--
You received this message because you are subscribed to the Google Groups "openSCADA" group.
To unsubscribe from this group and stop receiving emails from it, send an email to openscada+...@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

[Daniel B.]

Thanks Jens for the quick reply.

I've found several your answers about the same issue after send the issue, but it's not very clear for me what is the source of the problem OPC + DCOM itself, or j-interop, or Utgard?

Anyway we will perform some tests using standard OPC tools to see how Async groups works in case of DCOM connection.

четверг, 31 марта 2016 г., 20:32:49 UTC+3 пользователь Jens Reimann написал:

[Daniel B.]

Hi again Jens,

We finally got Utgard async groups to work over DCOM more or less stable. The problem solver is anyway:

JISystem.setJavaCoClassAutoCollection(false);

It does the trick. In spite of warnings are still there the updates continue to deliver from the OPC server after 20 minutes.

пятница, 1 апреля 2016 г., 9:59:32 UTC+3 пользователь Daniel B. написал: