Network skillset queue increasing with free agents presence

[ff7121]

Hello. Our customer got serious problem.

They have following configuration: 3 geographically distributed servers with installed Symposium CCS 5.0 (3 sites). These 3 sites are connected to each other via customer local network (CLAN). For call network routing between sites Symposium NCC 5.0 server is used. Each of Symposium server is connected to own Meridian station via embedded local network (ELAN).

The latest updates from manufacturer are installed on each server (SU_05, SUS_0501-SUS_0505). Each site has network skillset c_skillset that handles network routing.

Call routing script that runs on all sites looks as follows:

ASSIGN Prepaid TO c_Skillset
ASSIGN 4 TO c_Priority

IF (NOT OUT OF SERVICE c_Skillset) THEN
QUEUE TO SKILLSET c_Skillset WITH PRIORITY c_Priority
WAIT 2

QUEUE TO NETWORK SKILLSET c_Skillset WITH PRIORITY c_Priority
WAIT 2
ELSE
QUEUE TO SKILLSET Default_PrePaid WITH PRIORITY c_Priority
WAIT 2
END IF

SECTION Wait_Loop
IF (AGE OF CALL > MaxCallAge) THEN
DISCONNECT
END IF

IF (OUT OF SERVICE c_Skillset) THEN
DISCONNECT
END IF

IF (NOT QUEUED) THEN
IF (NOT OUT OF SERVICE c_Skillset) THEN
QUEUE TO SKILLSET c_Skillset WITH PRIORITY c_Priority
WAIT 2

QUEUE TO NETWORK SKILLSET c_Skillset WITH PRIORITY c_Priority
WAIT 2
ELSE
QUEUE TO SKILLSET Default_PrePaid WITH PRIORITY c_Priority
WAIT 2
END IF
END IF

WAIT 20

EXECUTE Wait_Loop

When CLAN works in a normal mode, system functionates stable with any system load.

If system load is low and a connection to one of the sites is dropped, following mistake occurs on other sites:

Severity: Critical
Code: 48117
Text: Interprocess Communications Error. (97, 0, 0, .\AsmCapi.cpp/NIasm_CancelAllRequests)

In this case other sites work stable and all services, shown with SMonW utility, are in UP state.

If system load is high and a connection to one of the sites is dropped, calls do not reach agent’s phonesets, pointed out above error occur and skillset c_skillset queue on working properly sites increases continuously. In the same moment of time free agents are present and a standard real-time report Standart_Agent_By_Supervisor shows, that non occupied agents are in RESERVED state.

The problem disappears after connection via CLAN is recovered or if manual data synchronization on NCC is made.

Such problem didn’t occur in similar situation with Symposium 4.2. According to problem description in official reference manual, Symposium server restart is needed. However system is installed at mobile operator of national scale, so Symposium server restart is possible only at night time. The situation looks critical and different from reference manual solution is needed. That’s why we addressed this problem to you.

 

[lretrievers]

Do you know which type of routing you are using (i.e. round robin, etc) and how many nodes you are targeting at a time, two or three?

CLAN traffic will effect your NCC and routing as it uses the CLAN to talk to the NCC server, and to check for available agents across the network.

 

[ff7121]

I clearly know how calls for network skillset works, because I setted up this system!

I am using round robin routing type.
Three nodes are registered on the NCC server at a time and network calls routing is present on each node.

This problem exposes itself only when the loading on all sites is high enough. When the load is low the problem doesnt appear.

If your a symposium guru and understand the core of the problem please give me more detailed answer.

 

[lretrievers]

There are a lot of factors that can come into the picture when you are trying to diagnose why there are problems in a networked environment. And there could be multiple reasons why this is occurring.

CLAN traffic is one, how many nodes you are targeting is another, timers is another and the list goes on. The agents staying in reserved when you loose the CLAN could happen since agents are reserved over the CLAN.

Do you know the reason for loosing CLAN connectivity?

 

[ff7121]

All 3 sites and NCC of CallCenter is located geographically remotely to each other and connected by communication equipment on CLAN. Periodically problems happen with communication channels and this problem shows up.

It is noticed that a problem shows up exactly at disappearing of connection with one of sites. At the loss of connection with NCC a problem does not show up.

I will repeat oneself, that a problem shows up only at a large load and at similar terms (temporal disappearing of connection between sites) the version of Symposium 4.2 worked correctly!

 

[ff7121]

Colleagues, maybe whoes have enyone other ideas? I really need yours help, because problem is critical for my customer!