cancel
Showing results for 
Search instead for 
Did you mean: 
cancel
8964
Views
20
Helpful
15
Replies

Intermittent One-Way Audio

msantiagodcny
Level 1
Level 1

We are experiencing intermittent One-way audio (in either direction) on both inbound and outbound off-network calls.

Our Deployement

CUCM: Ver. 8.5

Publisher:

Sub1: 

Sub2: 

Phones are only registered to either Sub1 or Sub2.

I have verified that the issue can occur regardless of which Sub the phone is registered to.

Two CUBEs - each with its own SIP trunk to our PSTN provider

OCIS-FR-3945-1  (no issue)

E911-FR-3945-1  (issue)

I have narrowed the issue down to calls which are going through our E911-FR-3945-1 CUBE and have had 0 reports of issues going out the other.

Phones:

Issue has been observed and reported against all of our deployed phone types: 7945, 7965, 7975.

I have been able to reproduce the issue about one in every 10-15 calls. I have had some success in tracing these problem calls using Wireshark to observe the SCCP call setup, and also the RTP streams.

To my surprise, on every observed one-way audio call that has been traced and saved, the audio packet streams are observable in both directions.

Furthermore, when a call is verified to have one-way audio during the call, compiling the RTP streams verify that the full two-way audio conversation is taking place on our network as I can hear both sides of it, only the affected end user does not hear it during the time the live call is taking place.

I can provide any necessary information as needed to help resolve this issue.

15 Replies 15

pkinane
Cisco Employee
Cisco Employee

What is the call flow?

It seems as though it is phone -> CUCM -> SIP trunk to cube -> ITSP

Where did you take the pcap from?

During the problem please get the following:

1. show call active voice brief from the CUBE along with calling/called numbers

2. show sccp connec from the CUBE

3. From the impacted phone, double press on help button '?' and check TX/RX packets are increasing or not.

Just recreated an Trouble call: one way audio.

Calling number: XXXXXX2911 (Us)

Called number: XXXXXX7574 (Them)

Calling Party could NOT hear Called party. 

pcap trace shows 3 rtp streams for some reason.

But both sides of the conversation are there even tho no inbound audio could be heard at the ip phone during the call.

1. show call active voice brief from the CUBE along with calling/called numbers

164C : 18547971 -1649868374ms.1 +6530 pid:101 Answer XXXXXX7574 active

 dur 00:06:11 tx:18554/2968640 rx:37107/5936980

 IP XXX.XX.100.164:13812 SRTP: off rtt:0ms pl:0/0ms lost:0/0/0 delay:0/0/0ms g711ulaw TextRelay: off

 media inactive detected:n media contrl rcvd:n/a timestamp:n/a

 long duration call detected:n long duration call duration:n/a timestamp:n/a

 

164C : 18547972 -1649868374ms.2 +6520 pid:200 Originate XXXXXX2911 active

 dur 00:06:11 tx:37107/5936980 rx:18554/2968640

 IP XXX.XX.209.166:16910 SRTP: off rtt:0ms pl:0/0ms lost:0/0/0 delay:0/0/0ms g711ulaw TextRelay: off

 media inactive detected:n media contrl rcvd:n/a timestamp:n/a

 long duration call detected:n long duration call duration:n/a timestamp:n/a

2. show sccp connec from the CUBE

E911-FR-3945-1#show sccp connections summary

SCCP Application Service(s) Statistics Summary:

Total Conferencing Sessions: 0, Connections: 0

Total Transcoding Sessions: 0, Connections: 0

Total MTP Sessions: 0, Connections: 0

Total ALG-Phone Sessions: 0, Connections: 0

Total BRI-Phone Sessions: 0, Connections: 0

Total SCCP Sessions: 0, Connections: 0

 

E911-FR-3945-1#show sccp connections

 

Total number of active session(s) 0, and connection(s) 0

3. From the impacted phone, double press on help button '?' and check TX/RX packets are increasing or not.

Rx/Tx frames were increasing under Network Statistics

Call Statistics showed fluctuating Rcvr Lost Packets: anywhere from -27360 to 27425

No Jitter

Avg MOS LQK: 2.0

Codec: G.711u

Can you upload the pcap file to look at the call?

phone -> CUCM -> SIP trunk to cube -> ITSP

This is the accurate call flow.

pcaps have been taken from the  PC port on the phone. 

Biggest surprise is that on trouble calls, the incoming audio packets are audible after trace, when they are not audible during the live call.

How are you taking the pcaps from the phone: SPAN to pc port in the phones configuration or using SPAN in the switch?

Since you are getting the pcap at the phone and you are able to play back both streams from the pcap, but can't hear the audio at the phone, I am curious about how many people are having 1 way audio at this particular site.

If they are using a headset, have you tested what would happen if you use the handset or speaker? I am willing to bet you will have the same situation due to you seeing high numbers for "Rcvr Lost Packets". It is an easy test to get out of the way though.

When you go to stream 1 on the phones web page during a bad call, take note of the remote IP address. Is it the IP address of the CUBE? If it isn't the CUBE, what is it?

Hello. I have same issue at one branch. Sometime once form 10-15 calls user have one way audio.

This issue even occur when call placed to neihbor phone (same subnet, same switch).

Is you resolve your issue?

Can you send the output of show proc cpu hist from the switch the phones use?

c2960X_st2_buz#show proc cpu hist 
      444444444444444444333334444433333444444444444444333334444433
      111222222222200000999991111199999000000000011111999991111188
  100                                                           
   90                                                           
   80                                                           
   70                                                           
   60                                                           
   50                                                           
   40 **********************************************************
   30 **********************************************************
   20 **********************************************************
   10 **********************************************************
     0....5....1....1....2....2....3....3....4....4....5....5....6
               0    5    0    5    0    5    0    5    0    5    0
               CPU% per second (last 60 seconds)

444444444444444444444444444444444444444444444444444444444444 244322322434354242244623242343244311432332332113233232226142 100 90 80 70 60 50 * * * 40 ########################################################## 30 ########################################################## 20 ########################################################## 10 ########################################################## 0....5....1....1....2....2....3....3....4....4....5....5....6 0 5 0 5 0 5 0 5 0 5 0 CPU% per minute (last 60 minutes) * = maximum CPU% # = average CPU% 444444644333333333333334444444544343333333333445545555655444444444544445 665677066788887778778874667578776909898987787382393524070978757969084883 100 90 80 70 60 * * * ** 50 ********* ********* ********************** * 40 #########**************##########**************##########************* 30 ###################################################################### 20 ###################################################################### 10 ###################################################################### 0....5....1....1....2....2....3....3....4....4....5....5....6....6....7.. 0 5 0 5 0 5 0 5 0 5 0 5 0 CPU% per hour (last 72 hours) * = maximum CPU% # = average CPU% c2960X_st2_buz#show proc cpu sorted 5sec | exclude 0.00 CPU utilization for five seconds: 43%/0%; one minute: 41%; five minutes: 40% PID Runtime(ms) Invoked uSecs 5Sec 1Min 5Min TTY Process 162 175051667 19924109 8785 24.76% 22.75% 22.65% 0 Hulc LED Process 228 1334830 5523208 241 1.01% 0.42% 0.38% 0 HULC DAI Process 131 6098349 797212 7649 0.59% 0.62% 0.60% 0 hpm counter proc 230 516288 5615370 91 0.35% 0.14% 0.12% 0 HRPC ip device t 17 1776355 3873380 458 0.35% 0.30% 0.33% 0 ARP Input 173 2868023 158825 18057 0.29% 0.30% 0.29% 0 HQM Stack Proces 240 268077 3507366 76 0.29% 0.08% 0.06% 0 IP Host Track Pr 168 250 144 1736 0.17% 0.24% 0.06% 1 SSH Process 174 962601 634186 1517 0.11% 0.11% 0.11% 0 HRPC qos request 92 331746 35882605 9 0.11% 0.10% 0.11% 0 RedEarth Rx Mana 220 781390 1722132 453 0.11% 0.06% 0.06% 0 Spanning Tree 84 85261 26651974 3 0.11% 0.04% 0.04% 0 Draught link sta 121 330321 21291819 15 0.11% 0.11% 0.12% 0 HLFM address lea 237 111615 7970754 14 0.05% 0.05% 0.05% 0 UDLD 207 90420 553169 163 0.05% 0.05% 0.02% 0 IP Input 183 994605 265738 3742 0.05% 0.12% 0.12% 0 Power RPS Proces 91 491696 35541570 13 0.05% 0.10% 0.11% 0 RedEarth Tx Mana

Two switch in stack 2960X. As I read early high cpu for Hulc LED Process its normal when use PoE powered IP phones.

Enable web access on both phones, then make the issue happen between two phones that are connected to the same switch if possible, when you reproduce the issue, keep the phone call active and browse to the web page for both phones. Once on the web page select stream 1 and document the remote address listed on the web page of phone A and the web page of phone B.

Here is an example of the page I am talking about:

https://supportforums.cisco.com/sites/default/files/legacy/5/5/3/90355-streaming1.jpg

Sorry, but screens with russian locale.

I can translate if something needed.

And I amazing, I never seen before codec OPUS. What is it?

All internal normal call always with g722 and external call always with g711.

Screen 1 translate:

Steam Status - Active

Sender Codec - No

Rcvr Codec - OPUS

Screen 2 translate:

Steam Status - Not Ready    - strange but 100% screenshot was make when call is active

Sender Codec - OPUS

Rcvr Codec - OPUS

I get call trace from RTMT 

In attacment two call.

I dont khow why but sip messanges is absent. I check this log first time atfer upgrade to 11.5 version. I check sericeability and trace enabled.

I check another failled calls from this branch, they always fail with 102 code. 

As seen on trace phone just not sent ACK on on OK from CUCM and after timeout CUCM finish call. This behaviour is same for all phones in this branch. 

Vlad,

Here are two links about OPUS:

http://opus-codec.org/

https://en.wikipedia.org/wiki/Opus_(audio_format)

Here is an enhancement request I filed not long ago. If you open a TAC case for this issue, request they attach this enhancement to the case.

https://quickview.cloudapps.cisco.com/quickview/bug/CSCvc39371

Please disable the OPUS codec on your system and test again. This can be done in the Service Parameters. Once you are on the Service Parameter page and you've selected the CallManager service, search for opus and disable it.

Thanks, disabling OPUS codec solve problem.