txjitavg and txjitmax have unreasonable values

[xaled1]

Hi,

after running the sipnagios I get the unreasonable values for txjitavg=97272.6ms txjitmax=481182.0ms more then actual call duration of 10 seconds
Also call disconnection shows Reason=603 (Decline) although it was disconnected from the sipnagios with BYE.

root@status:/usr/src/sipnagios/123/sipnagios# sipnagios --rtp-port=4000 --local-port=8090 --ip-addr=192.168.3.43 --local-user=+123456789 --local-siprealm=sip.test.com --local-password=cAcyAgaC46AKuRk sip:[email protected]:5060 --log-level=9 --app-log-level=9 -d 10
19:16:48.278 sipnagios.c !SIP UDP listening on 192.168.3.43:8090
19:16:48.279 sipnagios.c Making 1 calls to sip:[email protected]:5060..
19:16:50.346 sipnagios.c !....Call #0 connected in 2067 ms
19:17:00.359 sipnagios.c ....Call #0 disconnected. Reason=603 (Decline)
19:17:00.359 sipnagios.c ....Call #0 monitoring:
CRIT txjitavg=97272.6 MOS[rx=4.50 tx=4.50] Call To: sip:[email protected];tag=f22105bc TIME[pddms=47ms ttams=2067ms durationms=10014ms] RX[rxmos=4.50 rxpkts=501 rxbytes=80160B rxloss=0 rxlosspct=0.0% rxdup=0 rxduppct=0.0% rxreorder=0 rxreorderpct=0.0% rxjitmin=0.0ms rxjitavg=0.0ms rxjitmax=1.0ms] TX[txmos=4.50 txpkts=495 txbytes=79200B txloss=0 txlosspct=0.0% txdup=0 txduppct=0.0% txreorder=0 txreorderpct=0.0% txjitmin=0.0ms txjitavg=97272.6ms txjitmax=481182.0ms] RTT[rttmin=13.1ms rttmean=13.5ms rttmax=13.8ms] |durationms=10014ms pddms=47ms ttams=2067ms rxmos=4.50 rxpkts=501 rxbytes=80160B rxloss=0 rxlosspct=0.0% rxdup=0 rxduppct=0.0% rxreorder=0 rxreorderpct=0.0% rxjitmin=0.0ms rxjitavg=0.0ms rxjitmax=1.0ms txmos=4.50 txpkts=495 txbytes=79200B txloss=0 txlosspct=0.0% txdup=0 txduppct=0.0% txreorder=0 txreorderpct=0.0% txjitmin=0.0ms txjitavg=97272.6ms txjitmax=481182.0ms rttmin=13.1ms rttmean=13.5ms rttmax=13.8ms

[gmaruzz]

works for me, you want to try to understand what's wrong
get a pcap, or use sngrep to check it out

maybe you have ports that are not open, or natted?

maybe remote send you wrong RTCP?

try it from an host directly on the internet...

btw, I always use duration=40, maybe 10 sec is too short to accumulate enough RTCP pkts for stats
anyway, no idea :)

[xaled1]

Ok, will make some trace and see what will show up.

What about the "Also call disconnection shows Reason=603 (Decline) although it was disconnected from the sipnagios with BYE."?

[gmaruzz]

no idea :)
I would suspect has something to do with too short a call
maybe a failure code triggered by lack of data?

you may want to check it out what happened
test with duration=40 , if that way it works, try to understand why it fails with shorter duration

also, get a complete pcap and check it out in wireshark

I use sipnagios in production, with default values, from an host directly on Internet (eg: no NAT)

[xaled1]

One more thing - MOS calculation can not be correct rx=4.50 tx=4.50 with txjitavg=97272.6ms txjitmax=481182.0ms

It would also be good to be able to select audio codec using app parameter. Now PCMU is set by default. I had to change it in source code to PCMA to comply with European coding.

[xaled1]

I repeated the tests with default 40 seconds duration and still getting both issues:
Tests are running over internet, I just change the IP addresses when I post it to github

root@status:/usr/src/sipnagios/pjproject-2.11# sipnagios --rtp-port=4000 --local-port=8090 --ip-addr=1.2.3.4 --local-user=+123445677 --local-siprealm=sip.test.com --local-password=cAcyAgaC46AKuRk sip:[email protected]:5060 --log-level=9 --app-log-level=9
10:27:41.950 sipnagios.c !SIP UDP listening on 1.2.3.4:8090
10:27:41.950 sipnagios.c Making 1 calls to sip:[email protected]:5060..
10:27:44.012 sipnagios.c !....Call #0 connected in 2062 ms
10:28:24.018 sipnagios.c ....Call #0 disconnected. Reason=603 (Decline)
10:28:24.018 sipnagios.c ....Call #0 monitoring:
CRIT txjitavg=24318.3 MOS[rx=4.45 tx=4.50] Call To: sip:[email protected];tag=6dc6ff5d TIME[pddms=43ms ttams=2062ms durationms=40008ms] RX[rxmos=4.45 rxpkts=2000 rxbytes=320000B rxloss=1 rxlosspct=0.0% rxdup=0 rxduppct=0.0% rxreorder=0 rxreorderpct=0.0% rxjitmin=0.0ms rxjitavg=0.2ms rxjitmax=7.1ms] TX[txmos=4.50 txpkts=1995 txbytes=319200B txloss=0 txlosspct=0.0% txdup=0 txduppct=0.0% txreorder=0 txreorderpct=0.0% txjitmin=0.0ms txjitavg=24318.3ms txjitmax=481182.0ms] RTT[rttmin=8.7ms rttmean=9.4ms rttmax=10.5ms] |durationms=40008ms pddms=43ms ttams=2062ms rxmos=4.45 rxpkts=2000 rxbytes=320000B rxloss=1 rxlosspct=0.0% rxdup=0 rxduppct=0.0% rxreorder=0 rxreorderpct=0.0% rxjitmin=0.0ms rxjitavg=0.2ms rxjitmax=7.1ms txmos=4.50 txpkts=1995 txbytes=319200B txloss=0 txlosspct=0.0% txdup=0 txduppct=0.0% txreorder=0 txreorderpct=0.0% txjitmin=0.0ms txjitavg=24318.3ms txjitmax=481182.0ms rttmin=8.7ms rttmean=9.4ms rttmax=10.5ms

[gmaruzz]

¯_(ツ)_/¯
maybe it receives no/wrong rtcp from remote?

you may want to try to find out where those tx jit comes from, eg sprinkling printfs around

[gmaruzz]

(btw, you can raise the threshold for txjitavg CRIT and WARN, so sipnagios would still works for you even if remote is sending you wrong or no rtcp)

[xaled1]

I got it, nevertheless there is something wrong with tx MOS calculation, as it can never be 4.5 with txjitavg=24318.3ms

[gmaruzz]

no idea where those values comes from
please, check it out and send patches, if needed

[xaled1]

BTW what are you using as the B-side for the sipnagios?
I'm using asterisk that does some audio playback. Is it the correct way of operation?