CTI route points unregister when callmanager pub is shutdown

Answered Question
Jul 24th, 2010
User Badges:

UCCE with ICM 7.2.6 (proggers A and B) , IVR 5.0.2 (two IVRs), CCM 6.x ( one pub, 2 subs)


We notice that CTI route points unregister when the callmanager publisher is shutdown. Only one application that runs a IVR script stays registered. If subscribers are shutdown, no issue is seen with any CTI route points. What part of config am I missing? Any ideas?

Correct Answer by david.macias about 6 years 8 months ago

Well, yes because you have your pim pointed to the pub, why don't you try pointing it to a sub?


david

  • 1
  • 2
  • 3
  • 4
  • 5
Overall Rating: 5 (1 ratings)
Loading.
david.macias Sat, 07/24/2010 - 10:34
User Badges:
  • Blue, 1500 points or more

What do you see in your UCM PIM service?


david

khambetea Sat, 07/24/2010 - 19:11
User Badges:

I see messages like - "JTAPIGW received a message from CM that an address is out of service due to CTI Manager or CM failure. "

"ACD/IVR MGC_CCM (ID 5000) is off-line and not visible to the Peripheral Gateway. Routing to this site is impacted."

"TCP connection to JTAPI Gateway has been broken."


11:45:43 pg1A-jgw1 Trace: JVM Total Memory: 33423360 JVM Free Memory: 31528064 JVM Heap in Use: 1895296 JVM has 94.32943% free memory
11:45:43 pg1A-jgw1 Trace: CiscoAddrOutOfServiceEv Addr: 5074 Cause: CTIMANAGER_FAILURE
11:45:43 pg1A-jgw1 Trace:   MsgEMSEventReport:  MessageId: 2706702338 Arg1: 0 Arg2: 0 Arg3:  Arg4:  Arg5: 
11:45:43 pg1A-jgw1 Trace:   MsgAddressOutOfService:  Addr: 5074 Cause: 1007

---

11:45:43 pg1A-jgw1 Trace:   MsgAddressOutOfService:  Addr: 8739 Cause: 1007
11:45:43 pg1A-jgw1 Trace: CiscoAddrOutOfServiceEv Addr: 8493 Cause: CTIMANAGER_FAILURE

----

11:45:43 pg1A-jgw1 Trace: CallObservationEndedEv:  Ended Object: Addr: 1640
11:45:43 pg1A-jgw1 Trace: Encountered exception reading header: java.net.SocketException: socket closed
11:45:43 pg1A-jgw1 Trace: RecvHeader failed.
11:45:43 pg1A-jgw1 Trace: PIMConnection thread exiting

---

11:45:43 pg1A-jgw1 Trace: CallObservationEndedEv:  Ended Object: Addr: 2025
11:45:48 pg1A-jgw1 Trace: PIMServer: Accept connection only to loopbackaddress 127.0.0.1/127.0.0.1isloopback true
11:45:48 pg1A-jgw1 Trace: PIMServer: Accepted connection from 127.0.0.1/127.0.0.1
11:45:48 pg1A-jgw1 Trace: PIMServer: hostaddress 127.0.0.1/127.0.0.1
11:45:49 pg1A-jgw1 Trace: MsgOpenReq:  InvID: 358626645 Ver: 2 IdleTimeout: 80000
11:45:49 pg1A-jgw1 Trace: Initializing PIM Connection
11:45:49 pg1A-jgw1 Trace: Successfully initialized PIM Connection.
11:45:49 pg1A-jgw1 Trace:   MsgFailureConf:  InvID: 358626645 Status: 1
11:45:49 pg1A-jgw1 Trace: Encountered exception reading header: java.net.SocketException: Connection reset
11:45:49 pg1A-jgw1 Trace: RecvHeader failed.
11:45:49 pg1A-jgw1 Trace: PIMConnection thread exiting.
11:45:49 pg1A-jgw1 Trace: CancelRouteCallbacks from all DialedNumbers
11:45:49 pg1A-jgw1 Trace: JTapiObj: remove CallObservers from all addresses
11:45:49 pg1A-jgw1 Trace: JTapiObj: remove CallObservers from all addresses
11:46:52 pg1A-jgw1 Trace: PIMServer: Accept connection only to loopbackaddress 127.0.0.1/127.0.0.1isloopback true
11:46:52 pg1A-jgw1 Trace: PIMServer: Accepted connection from 127.0.0.1/127.0.0.1
11:46:52 pg1A-jgw1 Trace: PIMServer: hostaddress 127.0.0.1/127.0.0.1
11:46:53 pg1A-jgw1 Trace: MsgOpenReq:  InvID: 358689941 Ver: 2 IdleTimeout: 80000
11:46:53 pg1A-jgw1 Trace: Initializing PIM Connection


Looks like the jtapiservice registry key has only the publisher hostname. Should it have the subs as well - comma separated?

Correct Answer
david.macias Sat, 07/24/2010 - 21:26
User Badges:
  • Blue, 1500 points or more

Well, yes because you have your pim pointed to the pub, why don't you try pointing it to a sub?


david

pathyilbinu Sun, 05/22/2011 - 01:21
User Badges:

Dear David & all,

We have faced the same error in PGA side and as per david suggesion , we have noticed the A side Pim was pointed to Publisher and changed to subscriber but to no avail, recieved the same error (issue is that all agent logout all of a sudden and they can relogin and work)

Our one is a hosted instance and setup is as below version


UCCE with ICM 7.0 (PGA and B)

ICM7.0(0)_SR4

CTIOS7.0(0) with SR2 & ES13

Cisco CVP

CVP3.1(0),SR2 & ES3


CCM4.1(3)SR4d

( one pub, 4 subs)


Kindly advice as there is anything we can try out in config, will this error happen if the network is not stable ? the below error is happening frequency once in week or so ..


all the time i recived the error message like (either in A side PG or B side PG)  - "JTAPIGW received a message from CM that an address is out of service due to CTI Manager or CM failure. "

david.macias Sun, 05/22/2011 - 19:20
User Badges:
  • Blue, 1500 points or more

Yes, network problems will certainly cause this.  Can you move the PIM to another subscriber which might be in another site?  Get some information on your network and see if that's the issue.


david

pathyilbinu Mon, 05/23/2011 - 04:49
User Badges:

David,

Thanks for your prompt reply.

Are you telling the network issue between the Call Manager server and CTI server  ? already our other  instance is running fine with out any issue with the same config.(Pim is pointed to the same call manager sub for other instance and not reported any issue there)

Is there any other parameter we need to check ..thanks in advance.

The connection between ICM and CUCM is pure JTAPI.


Assume sub1 and sub2 are at the same site as JGW1A and sub3 and sub4 are at the same site as JGW1B. Assume JGW1A is configured to talk to sub1 and JGW1B is configured to talk to sub3. Assume JGW1A is active.


If you shutdown the CTI Manager process on the sub1 Call Manager, JGW1B will immediately go active and B will become the active PG. Intercluster signaling will take care of route requests on route points registered with sub1 - and they will come up the JTAPI connection on sub3 through the B side PG.


If you shutdown sub1 completely, those CTI route points will have to home onto sub3 but should do so somewhat seamlessly - in theory.


already our other  instance is running fine with out any issue with the same config.(Pim is pointed to the same call manager sub for other instance and not reported any issue there)


So there is another totally different ICM system pointing at your Call Managers using a totally different JTAPI user and this is not affected adversely by shutdowns?


Regards,

Geoff

pathyilbinu Tue, 05/24/2011 - 03:17
User Badges:

Thanks for your detail reply.

yes  Geoff, we do  have the hosted setup and  hence we have a different instance using different JTAPI user with the same setup (A & B sides) pointing to Call manager of A & B sites  and there we did not face any issue (all agent log out case) so far ..

Actions

This Discussion