Hi experts,

I have an use case where from OD I'm consuming a backend using a Java client. The backend might take till 30s to respond, and I know the VXML fetch timeout is 15s by default. If I increase the client timeout to 30s, I know the Session could be terminated and lost before the WS responds. So I'm analyzing options, what implication would increasing the fetch timeout to 35s have on the MPP/AAEP?

Another option could be to have retries on the client, and each retry to force a new VXML page to be sent to the MPP, so the fetch timeout timer would be refreshed on each new request to the backend.

I'm using OD 7.2.3, and AAEP 7.2.3 as well


Regards,
Jorge.

Hello experts,

We are currently working on a custom OD 7.2.3 app, to integrate against Dialog Flow. We are basing it on the DefautaDialogFlowApp and on the def_dialogflow.vxml. I would like to understand if there is a way to listen to Speech AND DTMF at the same time. AFAIK DTMF is processed by the MPP, and Speech by DF. The vxml is supporting both input modes at the same interaction, so I would like to understand how to do it using pure OD.

Thanks, Jorge.

Thank you for your answer Wilson.

We are currently using 07.23.08.01.

BTW, is there a way to change a call type once you have selected it? i.e. you want to change from Subdialog to Local refference. Once you drop the first refference to the module in your flow, you are asked to select Subdialog or Local, but from the 2nd time onwards, it always assume you want the call the same way as the 1st time.

Hello experts,

I have an OD 6.0 app (A) that invokes another OD project (B) as Subdialog. (B) receives input parameters from (A), executes, and then outputs other values.

When I migrated (A) & (B) to OD 7.2.3, the output values from (A) are coming wrapped between " ", hence any validation in (A) involving the return values from (B) fails. When I invoke (B) as Local, the output values don't contain " ", and (A) executes correctly.


Could you please let me know what could be causing the discrepancy?


Current Tech Specs:
Avaya Aura Orchestration Designer Version 7.2.3.0803
Apache Tomcat 8.5.45
AAEP 7.2.3

Previous Tech Specs:
Avaya Aura Orchestration Designer Version 06.00.10.02
Apache Tomcat 6
AAEP 6.0

Hi experts,

I have the following scenario that it's not working when I invoke an app as Local, but it does when I do it as Subdialog:

My main app (A) is calling another app (B)... B needs to throw a custom OD event named "Collection", and A needs to detect it locally (not in the AppRoot). So I put a Form node that throws "Collection" in B, and I placed a Catch for Collection in A where I call to B. The event is being catched only when I invoke B as Subdialog; if I do it as Local, the AvayaReturnErrorHandler is invoked and the event connection.disconnect.* thrown.


I need to know why it is not working when the app is called as Local.


Specs: OD 7.2.3, Java 1.8, Apache Tomcat 8.5

Thank you for your answer. What I am trying to understand is why am I getting the message:

"The WebLM URL configured in the DD Admin does not match the Cached WebLM URL used by DD applications in the runtime environment. You may need to restart the app server. "

As you pointed out, a value in the invalidlicensetimer tag would indicate a problem with the licenses or in the communication between the Application Server and the WebLM.

Before I imported the WebLM certificate into the AS' keystore, the log was saying "License is bad, allowing grace period". Once I imported it, I started getting:

[APP]|[04 Dec 2019 17:21:38,929]|[ INFO]|[CLASS: pruebaGoTo.writeln(57)]|[MESSAGE: BE0FE74E9EF7E42481FE4F95C781CFB4:/pruebaGoTo : Using license check timeout :0]
[APP]|[04 Dec 2019 17:21:42,132]|[ INFO]|[CLASS: pruebaGoTo.writeln(57)]|[MESSAGE: BE0FE74E9EF7E42481FE4F95C781CFB4:/pruebaGoTo : Runtime :true]
[APP]|[04 Dec 2019 17:21:42,132]|[ INFO]|[CLASS: pruebaGoTo.writeln(57)]|[MESSAGE: BE0FE74E9EF7E42481FE4F95C781CFB4:/pruebaGoTo : IC :true]
[APP]|[04 Dec 2019 17:21:42,132]|[ INFO]|[CLASS: pruebaGoTo.writeln(57)]|[MESSAGE: BE0FE74E9EF7E42481FE4F95C781CFB4:/pruebaGoTo : CTI :true]
[APP]|[04 Dec 2019 17:21:42,249]|[DEBUG]|[CLASS: pruebaGoTo.writeln(48)]|[MESSAGE: BE0FE74E9EF7E42481FE4F95C781CFB4:/pruebaGoTo : Grace Period timer cleared]
[APP]|[04 Dec 2019 17:21:42,250]|[DEBUG]|[CLASS: pruebaGoTo.writeln(48)]|[MESSAGE: BE0FE74E9EF7E42481FE4F95C781CFB4:/pruebaGoTo : License refresh timer started]
[APP]|[04 Dec 2019 17:21:42,252]|[ INFO]|[CLASS: pruebaGoTo.writeln(57)]|[MESSAGE: BE0FE74E9EF7E42481FE4F95C781CFB4:/pruebaGoTo : MessageType is speech]


So now the licenses seem to be fetched successfully, but I don't understand the reason of the message when I go into the validate jsp ("The WebLM URL configured in the DD Admin does not match the Cached WebLM URL used by DD applications in the runtime environment. You may need to restart the app server. ")

Hello all,

I'm having a problem with the WebLM license and the Application Server (Apache Tomcat 8.5). My IVR app is written in OD 7.2.1 and EPM is 7.2.3. Each time someone reboots the whole server, a new value is put in the invalidlicensetimer tag (at the ddconfig.xml file), resulting on "Invalid Runtime License" exception whenever a call is placed.

e.g.

<licenseserverurl value=""/>
<invalidlicensetimer value="1572555571410"/>
<licensechecktimeout value="0"/>
<proxy>

Each time I get this error, I need to shutdown the Tomcat, replace the value for an empty string (""), start it up again and then it works ok.

I have added the WebLM certificate into the .jks that the runtimeconfig is using, the License is being fetched ok now, but when I go to the validate page, I get:

"The WebLM URL configured in the DD Admin does not match the Cached WebLM URL used by DD applications in the runtime environment. You may need to restart the app server. "

I have restarted the app server several times, but the problem persists. I would really appreciate your help on this

Thank you.

This problem has been escalated with Avaya Support, apparently, no error in EPM... further investigations are going on in the MPP side.

Thank you Sir.

Actually this case was raised with CAS, they involved us, and then when I checked with the Speech Engineer I was pointed to the PH direction.

Do you know where could I open this to Avaya Support?

What do you think about the grammar? Do you think there's a way to make it more performant?

Experts,

The customer migrated from AAEP 7.0.2 to 7.2.0. The existing OD apps were migrated from 7.0 to 7.2.1.

In only 1 node, the customers are facing a silence 1-3 mins after hearing the next announcement node.

The grammar used by the node (8430_iaCheckingNumberDTMF.grxml) that is failing was not changed during the migration. It’s the same version that we delivered to the customer when we first deployed the app back in 2016 or so. We got no report of failure till nowadays.

I checked the logs, and the app just "stops" logging a random amount of time, and then it comes back to life. Checking the Voice Browser logs (there's no Nuance on this customer), I see no error or bad-fetch. Since the node collects sensitive data from the caller, it just shows some data as "private data".

If I change the node's grammar, for a fixed-length input, the node works as expected.

Timeouts are:

incompleteTimeout=1500
completeTimeout=0
terminatingTimeout=2000
interdigitTimeout=4000
timeout=3000


I'm attaching:
1) the grammar that's failing.... 2) another grammar also with variable amount of digits allowed that it's working as expected .... 3) the Voice Browser logs

I tested it entering 1-2-3-4-5-6-7-#, but with no luck.


Please let me know any additional detail you need,

Thanks in advance,
Jorge.

Environment details:

* Version: 7.0 (7.0.1.0804) Build id: 201410141416
* Apache Tomcat version 7.0.68
* AAEP version 7.0.2

Description

Customer reported error on their AAEP regarding a 404 Error code when fetching for an audio. I checked the AAEP & MPP logs and somehow OD is looking for the transitional audio (WSTransitionDelay) in the root folder of the od module webapp (for example, ONEGAS_GetAccount)


As I show in the screenshots attached, we have defined an Item Properties with the Transitional audio (WSTransitionDelay) in every AppRoot node of each OD module that contains Backend interactions. We have overwritten the hookGetPhrase method in the custom.java phraseset, in order to force OD to look for the transitional audio at a defined location (for example, <tomcatIP>:<port>/ONEGAS/Data/Audios/Custom).

The transitional audio seems to be working as expected, whenever you have a delay > 1sec, the audio starts playing. However, the reason why the AAEP errors are happening is unknown and I was hoping someone could bring a little bit of light into this problem.



Please let me know any additional information you might need to get the whole picture,

Thanks in advance,

Jorge.

Ok, will do that Ross.

Thanks!

Hi,

Our Spring Boot app DOES NOT use activemq-all library at all.

If we run our app in a clean Tomcat 8, it works fine.

If we add the runtime support files for Tomcat 8 in <tomcat_home>/lib (which includes the activemq-all library), our app fails due to the Exception presented above.

The problem is between spring classes inside the activemq-all, and our spring app libraries. That's why we try to determine if it could be exluded, or if there is a possible workaround.

Hi,

We are developing an application with spring-boot-actuator and spring-cloud. It's running fine in a Tomcat 8 with Java 8 without OD but the intention is to deploy it in a tomcat application server with some OD applications.

So, when we deploy the application in a tomcat server with OD dependencies, we are currently experiencing conflicts between spring classes inside the activemq-all library (which is present in the runtimeSupportTomcat8 ) and the spring classes being used by our application.

Here is the error:

//Exception encountered during context initialization - cancelling refresh attempt: org.springframework.beans.factory.BeanDefinitionStoreException: Failed to process import candidates for configuration class [org.springframework.boot.autoconfigure.transaction.TransactionAutoConfiguration$EnableTransactionManagementConfiguration$CglibAutoProxyConfiguration]; nested exception is java.lang.IllegalArgumentException: class org.springframework.transaction.annotation.TransactionManagementConfigurationSelector is not assignable to interface org.springframework.context.annotation.ImportSelector //


Java: 1.8
Tomcat: 8.0.46
JAR version: activemq-all-5.13.1
Spring boot: 1.5.4-Release
Spring core: 4.3.9-Release
Spring cloud zookeeper: 1.1.1.Release
OD: 7.20.09.04


If we remove the library activemq-all, our application works fine, so we would like to understand which is the activemq-all library purpose, what functionality is bringing to OD to figure out if it could be excluded or not.



Thanks in advance,
Jorge.

Hi,

I have faced an issue when setting up a new Prod environment. Here are the environment details:


Apache Tomcat 6.0.24
OD 07.00.18.01
Avaya Aura® Experience Portal 7.0
Java Oracle 1.7.0


When checking the tomcat's runtimeconfig, the keystore was configured as follows:


" Use other: /usr/lib/jvm/jre-1.7.0-Oracle.x86_64/lib/security/cacerts "

But all the Web services connections were failing due to a certificate issue. I enabled SSL debug mode, and could finally see that the actual trustore was being taken from "/usr/local/apache-tomcat-6.0.24/lib/trusted_weblm_certs.jks". This didn't make sense since the runtime config visual interface was showing it pointing to the cacerts file :(

I tried restarting it, and changing it several but with no results. Then I installed the WS certificates in trusted_weblm_certs.jks, and everything started working. Yet, the customer wanted the trustore to be the cacerts file, so I found a file inside the /lib in tomcat, named "trustedcert.properties" which content was:



WebLM.trustStorePath=relative
WebLM.trustStore=/trusted_weblm_certs.jks
WebLM.trustStorePassword=password

#WebLM.trustStorePath=absolute
#WebLM.trustStore=D:/winopenssl/OpenSSL/bin/demoCA/newcerts/weblmserver/trusted_weblm_certs.jks
#WebLM.trustStorePassword=password


I thought that the WebLM.trustStore option was the one in conflict with the configuration on the runtimeconfig, so I renamed the trustedcert.properties to trustedcert.properties.old. The app started to work ok and the trustore was now effectively taken from the cacerts file.

Approximately, one month later, the app started failing due a License issue, and because that trustedcert.properties file was needed for something (OD was working because of the grace period). I renamed it back to the original form. Everything started to work normally again, and the license was now taken from the EP WebLM configured Server. And the keystore is still taken from the cacerts file :)


I would like to ask:

1. Which is the function of the trustedcert.properties file.

2. What could have caused that difference between what the runtimeconfig was showing as the keystore (/usr/lib/jvm/jre-1.7.0-oracle.x86_64/lib/security/cacerts), and the actual (/usr/local/apache-tomcat-6.0.24/lib/trusted_weblm_certs.jks).

3. If you do not set a License Server on the runtimeconfig, it will use the License Server configured on the EP automatically, right?

Thanks in advance!