Hi Robert,
1. as written - the long form is the only one possible in many cases;
where both can be used, the short one is often preferred by
keyboard-savvy (or lazy, like me) people.
2. where has the authentication part (INVITE, 100, 401, 407) gone? Not
that its absence would explain why there is no response from the VM, it
just differs from the scenario as you've shown it initially.
3. if Wireshark explicitly says that the UDP checksum is correct in the
packet dissection pane (not just that it doesn't show the packet in
red), then it is possible that your version of tcpdump has an issue with
UDP checksum calculation. Otherwise, UDP checksum evaluation is often
switched off by default in Wireshark. Go
Edit->Preferences->Protocols->UDP and check whether the "validate UDP
checksum if possible" is ticked.
4. looking at your /etc/hosts, I don't get how your tcpdump translates
129.82.3.26 to localhost.localdomain, but the source address of the
packets seems to be OK, so the UDP checksum is more likely to be the
reason why the voicemail does not respond (I suppose that there is no
firewall and that the voicemail does listen at UDP/5060).
P.
Dne 4.8.2016 v 22:03 Remsik,Robert napsal(a):
1. Ah, that helps explain it! Thank you. 😊 Is the long form or short
form generally preferred?
2. Thank you, that did the trick. It now sees the entire scenario.
Messages Retrans Timeout
Unexpected-Msg
INVITE ----------> 1 5 1
100 <---------- 0 0 0 0
180 <---------- 0 0 0 0
183 <---------- 0 0 0 0
200 <---------- E-RTD1 0 0 0 0
ACK ----------> 0 0
Pause [ 0ms] 0 0
BYE ----------> 0 0 0
200 <---------- 0 0 0 0
3. Interesting... I'll sniff traffic as it exits towards voicemail
and verify if it detects a checksum error.... No checksum errors
showing up in Wireshark. I'll try to comb through some more logs to
see if I can turn anything up.
4.
$ grep localhost /etc/hosts
127.0.0.1 localhost localhost.localdomain localhost4
localhost4.localdomain4
::1 localhost localhost.localdomain localhost6
localhost6.localdomain6
13:40:50.338851 IP (tos 0x0, ttl 64, id 14685, offset 0, flags [DF],
proto UDP (17), length 551)
129.82.3.26.sip > 10.20.128.149.sip: [bad udp cksum 0x113a ->
0x15d2!] SIP, length: 523
INVITE sip:15500@10.20.128.149:5060 SIP/2.0
Via: SIP/2.0/UDP 129.82.3.26:5060;branch=z9hG4bK-21667-1-0
From: sipp <sip:sipp@129.82.3.26:5060>;tag=21667SIPpTag001
To: 15500 <sip:15500@10.20.128.149:5060>
Call-ID: 1-21667@129.82.3.26
CSeq: 1 INVITE
Contact: sip:sipp@129.82.3.26:5060
Max-Forwards: 70
Subject: Performance Test
Content-Type: application/sdp
Content-Length: 133
v=0
o=user1 53655765 2353687637 IN IP4 129.82.3.26
s=-
c=IN IP4 129.82.3.26
t=0 0
m=audio 6000 RTP/AVP 0
a=rtpmap:0 PCMU/8000
Robert Remsik
ACNS
Desk Phone: 970 491 7120
robert.rem...@colostate.edu
------------------------------------------------------------------------
*From:* sindelka <sinde...@ttc.cz>
*Sent:* Thursday, August 4, 2016 1:09 PM
*To:* Remsik,Robert; sipp-users@lists.sourceforge.net
*Subject:* Re: [Sipp-users] Troubleshooting a basic call
Robert,
1. adding </recv> after my recv tags which have already been closed by
the / right before the closing > was not the right thing to do. SIPp
is not really good in detecting wrong syntax in the scenario file,
which explains why only the initial INVITE and 100 are displayed on
the runtime screen now after your modification.
In fact,
<recv attribute1="value1" attribute2="value2"*/*>
is equivalent to
<recv attribute1="value1" attribute2="value2">
</recv>
but shorter. Of course, the longer variant is the only possible one
for tags which contain a body, like all cases of send or if the recv
contains some action, like ereg.
2. the actual mistake causing SIPp to complain about an undefined
label was a typo: a "-" (dash) instead of "_" (underscore) in the
label name (after the <recv response="407"/>)
3. the bad UDP checksum is likely to cause the voicemail server not to
answer, but I'm not sure it is a SIPp issue - normally, an application
sends only the payload to a protocol socket, and the calculation of
checksums is a job of the protocol stack in the kernel. If your
network card eventually supports UDP checksum offloading, it may be
that the error doesn't actually exist, although in such case I would
expect the wrong value shown by tcpdump to be 0. To check that, you'll
need to capture the packets anywhere else at their way from the sender
to the recipient but at the sending machine (because there, the
capturing point is before the UDP checksum is calculated if the
offloading is used). If you can see the UDP checksum error after the
packet has left the sending machine, it is really there and likely to
cause the receiving machine's IP stack to drop the packet rather than
deliver it to the application.
4. can you please run the tcpdump with -n option added, or place the
output ofgrep localhost /etc/hosts here?
It seems strange to me that the packet would have 127.0.0.1 as source
address as the localhost.localdomain fqdn suggests, but if it does,
the voicemail might to send the responses to itself if it looks at the
source address of the request packet rather than at the address given
in the Via header.
P.
Dne 4.8.2016 v 20:34 Remsik,Robert napsal(a):
Thank you for the help! This is indeed my base case copy+pasted,
just calling into voicemail and then either voicemail hangs up or I
hang up.
I added the modifications you suggested but I had to add </recv>
after each recv tag to have the program not throw an error of: (The
label 'auth_challenge_received' was not defined (index 2, next
attribute)). However now the scenario only lists an invite and 100
message, not the full scenario that I was expecting.
Interestingly the INVITE lists a bad UDP checksum. That might help
explain why I'm not getting a response back from anything.
12:35:17.894298 IP (tos 0x0, ttl 64, id 62100, offset 0, flags [DF],
proto UDP (17), length 891)
localhost.localdomain.sip > 10.20.128.149.sip: [*bad udp cksum
0x128e -> 0xb0db!*] SIP, length: 863
INVITE sip:17120@10.20.128.149;user=phone;transport=tcp SIP/2.0
Via: SIP/2.0/UDP 129.82.3.26;branch=z9hG4bK-17371-1-0
From: "Robert Remsik" <sip:17...@otc.colostate.edu>;tag=1
To: <sip:17120@10.20.128.149;user=phone>
CSeq: 1 INVITE
Call-ID: 1-17371@129.82.3.26
Contact: <sip:17120@129.82.3.26;transport=UDP>
Allow: INVITE, ACK, BYE, CANCEL, OPTIONS, INFO, MESSAGE,
SUBSCRIBE, NOTIFY, PRACK, UPDATE, REFER
User-Agent: PolycomVVX-VVX_410-UA/4.1.8.0628
Accept-Language: en
Supported: 100rel,replaces
Allow-Events: conference,talk,hold
Max-Forwards: 70
Content-Type: application/sdp
Content-Length: 245
v=0
o=- 1470166919 1470166919 IN IP4 129.82.3.26
s=Polycom IP Phone
c=IN IP4 129.82.3.26
t=0 0
a=sendrecv
m=audio 6000 RTP/AVP 9 0 8 101
a=rtpmap:9 G722/8000
a=rtpmap:0 PCMU/8000
a=rtpmap:8 PCMA/8000
a=rtpmap:101 telephone-event/8000
Messages Retrans Timeout
Unexpected-Msg
INVITE ----------> 1 0
100 <---------- 0 0 0 0
<recv response="100" optional="true"/>
</recv>
<recv response="401" optional="true" auth="true"
next="auth_challenge_received"/>
</recv>
<recv response="407" auth="true"/>
</recv>
<label id="auth_challenge-received"/>
</recv>
Robert Remsik
ACNS
Desk Phone: 970 491 7120
robert.rem...@colostate.edu
------------------------------------------------------------------------
*From:* sindelka <sinde...@ttc.cz>
*Sent:* Tuesday, August 2, 2016 9:12 PM
*To:* sipp-users@lists.sourceforge.net
*Subject:* Re: [Sipp-users] Troubleshooting a basic call
Hi Robert,
if the scenario you've provided is a verbatim copy of the one you
actually use, not an edited version for illustration purposes, the
trouble is that you blindly fire a series of requests without waiting
for any response from the server. Normally, the authentication challenge
comes in a 401 or 407 response to the first INVITE, and the generation
of the authentication response as requested by the [authentication]
keyword depends on the information from the authentication challenge.
So you must insert some optional and mandatory <recv> statements between
the first INVITE and first ACK - namely,
<recv response="100" optional="true"/>
<recv response="401" optional="true" auth="true"
next="auth_challenge_received"/>
<recv response="407" auth="true"/>
<label id="auth_challenge-received"/>
This ensures that regardless whether the sipx device sends the 100 or
not, and regardless whether it uses 401 or 407 to send the
authentication challenge, you'll handle it properly.
Also, remove the auth="true" AVPs from all the <send> blocks, they are
useless there.
On top of that, please <recv> optionally a 100, a 180, a 183 and then
compulsorily a 200 between sending the second INVITE (with the
authentication response) and sending the second ACK, otherwise you may
run into some other issues.
And the last point, if you specify user and password inside the
scenario, SIPp does not use the -au and -ap specified on the command
line.
Pavel
------------------------------------------------------------------------------
_______________________________________________
Sipp-users mailing list
Sipp-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/sipp-users
Sipp-users Info Page - SourceForge
<https://lists.sourceforge.net/lists/listinfo/sipp-users>
lists.sourceforge.net
This is the sipp users mailing list. Use it to get support for sipp
or ask for features or even post your own features to be included in
sipp!
------------------------------------------------------------------------------
_______________________________________________
Sipp-users mailing list
Sipp-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/sipp-users
