IPCC Enterprise Failover

Unanswered Question
May 12th, 2008
User Badges:

We are testing with an IPCC Enterprise 6.0 failover cases. We have tried a number of failover cases and 2 of them are quite strange.

1. Sub CCM down during an agent answering an ACD calls. The agent dekstop (CTIOS toolkit) will be logged out and agent need to log-in again to CTIOS.

2. When CG / CTIOS service down on 1 side during an agent answering an ACD calls. The agent desktop has tried to re-login but in "Ready" state while agent is still answering the previous ACD calls.

Is this a expected behavior from IPCCE side and has anyone share similar testing result to us.


  • 1
  • 2
  • 3
  • 4
  • 5
Overall Rating: 0 (0 ratings)

Assuming you have a publisher and two subscribers, the JTAPI connection on PG1A will be to sub1 and the JTAPI connection on PG1B will be to sub2.

Let's say PG1A is the active side. If you stop the CTI Manager process on sub1, the idle side will immediately go active to sub2. Phones that are homed onto sub1 will not move (will not home to sub2), but intracluster signalling will enable CTI events for phones on sub1 to be passed to the CTI Manager running on sub2 and then up to ICM via JTAPI.

If you shutdown CCM sub1, phones homed to that sub will switch over to sub2 - but the case is the same.

Do you have the PIMs pointing to the different subs as described? This is detailed in the SRND.

When you have the two CTIOS configured correctly, and the desktop is pointing to A and B, should one be stopped the other will provide the feed.

Your current behaviour is not demonstrating the fault tolerance we would expect to see.



ibm.alexl Tue, 05/13/2008 - 17:57
User Badges:

We are actually going in a POC stage with 1 pub and 1 sub.

I have checked that CCM-PG-A is pointing to PUB1 and CCM-PG-B is pointing to SUB1. Both side should have the CTI-Manager service activiated. Should it makes any difference in compare to using 2 Sub for CCM-PG?

No, that should not matter.

Can you check the JTAPI gateway on each side when you stop the other side?

Say side A jgwy is active. Watch jgwy on side B - it will say idle. Restart the PG on side A and as it goes down, the jgwy on side B should immediately go active. Then flip over and watch the jgwy on side A - should say idle, and immediately go active when you restart the PG on side B.



ibm.alexl Fri, 05/16/2008 - 01:44
User Badges:

Yes. I have checked the jgwy process and that one should be active and one should be idle.

I tried to cycle the CCM PG and i think it's OK for the failover. But when cycle the CG / CTIOS server. The CTI status is not maintained when agent is in a call so that the failover CTIOS server is treating the agent to be ready and receive calls.


This Discussion