Core Issue
This is the topology:
IP PHONE 1 IP PHONE 2
| |
CCM SUB --------------CCM PUB |
CCM CLUSTER
|------------------------------------ |
PSTN GW 1 PSTN GW 2
IP phone 1 is registered to the Cisco CallManager publisher and calls to the plain switched telephone network (PSTN). The Cisco CallManager picks up H323 PSTN GW 1 as the primary gateway due to the device pool setting to route the call out from the Cisco CallManager subscriber through the H.323. These calls drop intermittently during mid-call and when the call is dropped, the IP phone displays a Temp Fail error message.
Resolution
The Temp Fail message appears when the phones unregister from the Cisco CallManager server to which they connect. This can be caused by network issues that prevent the phones to connect to their Cisco CallManager server or maybe the box they register to went down and the phones failed over to the next server. Application Log of the Event Viewer on the Cisco CallManager server the phones are registered to show some related errors.
This is a sample trace captured on this issue:
- The Cisco CallManager instructed the IP phone to drop the call:
03/13/2006 10:13:18.855 CCM|StationD: (0008741) DisplayPromptStatus timeOut=0 Status='#' content='Temporary failure' line=1 CI=17763918 ver=84000006.|<clid::standalonecluster> <nid::192.168.56.2><ct::0><ip::><dev::>
- When the call is dropped, the Cisco CallManager trace shows an SDL link between PUB and SUB as out of service:
03/13/2006 10:13:18.730 CCM|MediaExchange(658803) - SdlLinkOOS: Received SdlLinkOOS for nodeID 2. (PartyA nodeID=1, PartyB nodeID=2)/n|<CLID::StandAloneCluster><NID::192.168.56.2><CT::0,0,0,0 .0 > <IP::><DEV::>
- The SDL trace reports SDL link out of service as well.
153030980| 06/03/13 10:13:18.684| 001|LnkState | <closing connection> Reason: Read Error-errno:10054, NodeId: 2, AppId: 100, TCPAddr: 192.168.56.3:50930 153030981| 06/03/13 10:13:18.684| 001| LnkState | <linkoos> NodeId: 2 AppId: 100 [192.168.56.3:8002]
If traces show similar symtoms as previously shown, then:
- Check the network path between the Subscriber and Publisher Cisco CallManager Servers.
- Check the switch ports on which the Cisco CallManager Server is connected.
If you receive errors on switch ports, change the NIC setting on the server. If the NIC port speed is hardcoded to 100/Full, change it to 100/Auto. This can help clean the error, which is mostly caused due to port setting mismatch.
When no errors are seen on switchport of the switch, the SDLLINK out-of-service message should no longer be reported, and the call drops so the issue should be resolved.
Problem Type
Call control software (CallManager, CallManager Express, ICS7750, SRST, SS7 call agents)
IP phones and other clients (Client registration, faulty IP phone, etc.)
Call Control
CallManager
Call Connection
Dropped or disconnected calls
CallManager Features
Bad operational behavior, status,information, performance, delay or other CCM failure
IP Phones and Clients
IP phone error messages
Failure Type
Drops out, times out, flaps, toggles, hangs, gets stuck, locks up, loops, temporary pause, stops responding, cuts short, etc.
Error message
Cannot connect to server, disconnect, register, synchronize, communicate, etc.
Frequency and Scope of Symptom
Occurs intermittently
When Problem Occurs
During conversation
End Where Problem is Heard or Device Type Involved
IP phone
Call Topology and Direction
IP network > PSTN/PBX/gateway
IP Phone Errors
Temporary failure
MGCP Features
MGCP gateway
CallManager Versions
CallManager 4.x
CallManager 3.x
Computer Telephony Integration (CTI) Port
Gateway device
CallManager System Features
Device pool
CallManager Distributed Call Processing
CallManager cluster