We have some calls disconnected during the conversation. These failures happen random on different phones and users several times a day.
The phones are SIP Ip communicators.
We have CUCM cluster
10.13.81.7(sub1 - the server was shut down)
10.13.81.16(sub2 - the only call-processing node )
I found this error messages at the log file(call from 934760460 to 8267 at 15:17 )
Thu Feb 12 15:17:10.447 : DET : ( 4984) -VM| main|cip.app.bv:? - reservationGranted - render ( "From 934760460" )
Thu Feb 12 15:17:10.447 : EE : ( 4984) main|cip.app.bv:? - reservationGranted - render ( "From 934760460" )
Thu Feb 12 15:17:12.467 : EE : ( 3308) -SFB| --> CSoftFireBoxWnd::onKeyUp()
Thu Feb 12 15:17:12.467 : SPCL : ( 3308) -SFB| &S CSoftFireBoxWnd::onKeyUp(): wParam_ = <44>, lParam_ = <0xC1370001>
Thu Feb 12 15:17:12.467 : EE : ( 3308) -SFB| <-- CSoftFireBoxWnd::onKeyUp()
Thu Feb 12 15:17:21.697 : ERROR : ( 4116) SIP : sip_tcp_create_connection : socket connect failed errno: 34 ipaddr: 10.13.81.7, port: 5060
Thu Feb 12 15:17:21.697 : ERROR : ( 4116) SIP : sip_tcp_detach_socket : Max TCP connections reached.
Thu Feb 12 15:17:21.697 : DET : ( 4116) SIPCC-SIP_TCP_MSG: sip_tcp_purge_entry: Socket fd: 1396 closed for connid 1 with address: 1, remote port: 168644871
Thu Feb 12 15:17:21.697 : ERROR : ( 4116) sip_regmgr_ev_tmr_ack_retry: Error: sip_platform_tcp_channel_create(server addr=10.13.81.7, server port=5060) failed.
Thu Feb 12 15:17:21.697 : DET : ( 4116) SIPCC-SIP_FALLBACK: 60/1, sip_regmgr_retry_timer_start: Starting fallback timer (120 sec)
Does CUCM have tcp connections limit? is it configurable? Or maybe it is windows machine's error?
Attached full log
To me it seems like IPC lost communication with CUCM. You mention that one of the subscribers are shut down, are you planning to leave it this way? If so, its recommended to remove it from the cluster else you will have replication issues.
Yes, we are planning to back sub1 to the cluster as soon as server is fixed after hardware failure. But I was reported that the problem with phones existed before sub1 failure.
How can I prove that this is a network failure?
TY for your help. Appreciate it.