Michal,

Very cool. Thanks for the reply.

I was ready to work around this by invoking multiple web-services but I am going to give this POJO approach a shot.

Thanks!

-Ian

Hi Michal,

Were you able to find an answer to this question? If so, can you please let me know?

Thanks,

-Ian

Hi Douglas.

We had the same issue.

ref: https://devconnect.avaya.com/public/forum/d_forum_3.jsp?c=0&f=6&t=13196

-Ian

Here is an update on the issue.

We have not seen any of the same latency (have a different issue for the 1st call of the day, but that appears to be different).

Big thanks to you Wilson! Your help was much appreciated!

-Ian

Just want to give an update on this one.

We were able to find a solution. We edited the grammar in text mode and manually set the builtin type to 'touchtone'.

<?xml version="1.0" encoding="UTF-8" standalone="no"?> <grammar>
<properties>
<mode value="dtmf"/>
<type value="static"/>
</properties>
<data>
<options value="minlength=1;maxlength=10"/>
<builtintype value="touchtone"/>
</data>
</grammar>

We have implemented the username+password recommendation. Thanks Wilson. At first glance the transactions appear to be quicker.

We will know later on in the day whether this has fixed our issue.

I will post an update at the end of the day.

Thanks.

The code that I had posted was code that Oracle had given us to test in a separate java servlet application. They had asked us to use this code to validate the connection between the WebLogic and the DB. The result is that we haven't seen any delays from this code. Only the OD code. That is why the investigation has shifted from WebLogic, to figuring out what is happening with OD/app.

Our OD application uses the DBOP in its standard format. We have not added any extra coding to the JAVA.

In the sample code that you had sent me, the initContext had already been defined. How is it defined? You can see from the code that Oracle gave us that the Initial Context was very specific.

Is there any extra logging that we can enable in order to better understanding the issue? We have turned the standard logging on but we do not see any of the problems related to gathering a datasource (if they exist).

Thanks. -Ian

How is the initContext built?

In the sample code that Oracle gave us, we used:

Context ctx = null;
Hashtable ht = new Hashtable();
ht.put(Context.INITIAL_CONTEXT_FACTORY, "weblogic.jndi.WLInitialContextFactory");
ht.put(Context.PROVIDER_URL, "t3://hostname:port");
ctx = new InitialContext(ht);

Oracle have looked at the issue and they have asked to see the java code, which I can't provide because my OD has the DBOP encapsulated.

They would like to know how the JDBC request is being put together, are there any delays, etc.

The data source name in WebLogic has been changed to be prefixed with 'jdbc/' and that has made a big difference (Thanks Willson Yu!!!).

The problem has morphed though. Most of the time we are getting ms response times however during some 10 minutes (or so) stretches of the day, we get 7 second response times. We don't understand why. The server does not appear to be in any sort of CPU trouble at the time and neither does the DB.

When we look at the JDBC logs, we can see the 7 second delay, however we do not understand what is happening during the delay.

<8-May-2013 2:46:06 o'clock PM EDT> <Debug> <JNDI> <BEA-000000> <+++ lookup (jdbc.iandb, weblogic.rmi.cluster.ClusterableRemoteObject) succeeded>
<8-May-2013 2:46:14 o'clock PM EDT> <Debug> <JDBCConn> <BEA-000000> <JTS/JDBC Connect: url = jdbc:weblogic:jts:jdbc/iandb, tx = null, props = {EmulateTwoPhaseCommit=false, connectionPoolID=jdbc/iandb, jdbcTxDataSource=true, LoggingLastResource=false, dataSourceName=jdbc/iandb}>


Any suggestions or comments would be greatly appreciated.

Thanks,

-Ian

Hi Deepak. Did you figure out what the issue was? I am interested to know. Thanks.