Fixing One-X Application most common issues

Learn to tackle everyday issues by applying these 4 procedures.

By the end of this post, “Fixing One-X Application most common issues,” you have learned how to identify issues related to Call Handling, determine problems involving One-X Application registrations, and finally see how to avoid conflicts that can cause the application to freeze, if not configured correctly.

A customer states the One-X Application is freezing whenever they try to answer calls and perform routine call handling functions. Another customer called saying that he is no longer able to register to his H.323 One-X Agent Application, and a third person complained about not able to hear while calls were originated from the One-X Communicator.

As these are some of the regular issues that you might encounter, I decided to provide you the following 4 procedures to help you find fixes or help you along the way to fix them.

• 1.- Investigate and Document
• 2.- Learn the environment
• 3.- Research and test
• 4.- Develop a solution

1.- Investigate and Document

While performing the data gathering, try to focus on the customer’s experience. Ask what type of errors or messages they experience or see whenever using the One-X App.

Schedule a meeting where you are able to see the end-user desktop, while they are experiencing the same problems.

Screenshots are also very helpful, as these help you visualize the app’s behavior.

One-x Application Hang-ups and Call Handling Issues

This is a common misconfiguration problem that most of the System Administrator do whenever configuring the One-X Agent, leaving the Auto Answer and other Application features set as default. As you document these type of events, it is important to recognize how the customer wants these featured configured and how will they want the calls handled once presented to the Agents.

One-x Application not registering

Whether you are using PROCR or CLAN registration, you have to understand the process each end-user is taking to communicate with these registration servers.

As you troubleshoot these type of issues, have the user come to the office instead of connecting from remote location, if possible, to eliminate any VPN connection issues.

One-x Application not able to hear while on a call

These are normally events that happen to remote workers. As you work on gathering the details, ask to see if anything changed in the network environment. You are looking to document anything from updated technologies, new implementations, rules, etc.

2.- Learn the environment

Now that you have a good understanding of the network environment, it is time for you to start seeing how these components talk to each other.

Call Handling Issues and their environment – Whether these are users located at the HQ office or working remotely, it is important to know which resources are utilized whenever they are receiving a call and performing any call handling.

You are looking for Locations, IP Network Regions, Network Maps, and type of Media Processors.

Registration – For those using CLAN for registration, I recommend to use PROCR to capture the registration events from CM as CLANs are not able to report to a local directory, unless a SPAN/Mirror port is setup to capture network events to a PCAP file using a Network Analyzer.

Audio – Similar to registration events, you have to capture the network events to confirm the RTP Stream offer is been sent to the One-x.

3.- Research and test

By now you should know what the problem is. For a quick solution, head over to the http://support.avaya.com to see if there are any know PSNs created answering your problems.

For those who are not so lucky, lets create the following testing procedure=

Call Handling

As Auto Answer is the main conflicting issue here, you need to decide if the One-X’s user is manually answering the calls, or if you prefer having CM triggering the Auto Answer instead.

Manual Auto Answer

1. In the Station Form, second page Auto Answer= None
2. One-X Agent Application – CM Auto Answer Support Required= is unchecked
3. One-X Agent Application – Work Handling= Auto Accept
4. CM Performing the Auto Answer

CM Agent-ID Form – Auto Answer set to= Station

1. In the CM Station Form – second page AA= ACD or All to Auto Answer all station and ACD calls.
2. One-X Agent Application – CM Auto Answer Support required= is checked
3. One-X Agent Application – Work Handling= Manual Accept

Registration

The registration uses UDP 1719 with TCP 1720 to imitate the registration process with dynamic UDP/TCP port ranges from 1024 to 65535.

The following are the registration steps taken from the One-x application when using PROCR as the registration server=

• PROCR sends a “Gatekeeper Request” message over RAS/H.225, the receives a “Gatekeeper Confirm”, followed by a couple of “Non Standard Messages” interchanged between both One-x Application and PROCR.
• TCP is used to send SYN/ACK Messages to complete the registration process. By default the CM sends the last registration message back to the One-x Application.
• If the last TCP message is failing to get back to the Application, you can force the Application to send this messages instead by disabling the TCP Signaling Socket option on the IP Network Region (2nd Page) associated with this particular One-x Agent/Communicator Application.
• Besides capturing these events through Wireshark (tshark/dumpcap/tethereal), you can check the the log manager “/var/log/ecs” and messages to see the registration failures.

Audio

These UDP ports are defined in the IP-Network-Region first page (UDP Port minimum/Port Maximum) options. They are pre-set ports sent from CM out to One-X Agent. The One-x Agent uses a range of dynamic port listed above, which can also be defined within the One-X Application.

It is always important to have the network admin review any rules, VPN Profiles, Access Control Lists, and H.323 Deep Packet Inspection is not enabled.

4.- Develop a solution

After testing the different scenarios, it is time to implement the new-found fix. Depending on the amount of users, you might schedule a batch of users first to see how these operate, before performing a massive change.

 

Question – When t-shooting these type of issues, what is the first step you take to fix it?