Update: Disabling the topoh module on the proxy which produces the error
seems to stop the failure from manifesting. I'll try using topos_redis
instead, but should this be treated as a bug?

BR,
George

On Wed, 8 Jul 2020 at 12:37, George Diamantopoulos <[email protected]>
wrote:

> Hello again,
>
> Indeed $mb seems to contain garbage:
>
> SCRIPT_MB: ACK <BC><EA><8F> SIP/2.0#015#012Via: SIP/2.0/UDP
> 172.30.154.189;TH=dcv;branch=z9hG4bK629b.6af9302cd78dc58dffe817e60124f4ed.0#015#012Route:
> <sip:[email protected]
> ;lr;received=sip:2.2.2.2:32768
> ;ob;r2=on>,<sip:[email protected]
> ;lr;received=sip:2.2.2.2:32768;ob;r2=on>#015#012Max-Forwards:
> 68#015#012From: "Anonymous" 
> <sip:[email protected]>;tag=as4bc9e324#015#012To:
> <sip:[email protected]>;tag=jw7z5s0zvc#015#012Call-ID:
> [email protected]#015#012CSeq
> <http://[email protected]#015%23012CSeq>:
> 102 ACK#015#012Content-Length: 0#015#012TH: dch#015#012#015#012
>
> How can this be possible? Capturing traffic on wire shows the RURI I
> pasted in my original message and there are no script operations on the
> RURI before sanity_check() (message buffer above is printed just before
> sanity_check() is run in REQINIT).
>
> BR,
> George
>
> On Wed, 8 Jul 2020 at 11:18, George Diamantopoulos <[email protected]>
> wrote:
>
>> I'll post the message buffer ASAP, but in the meantime I don't see how
>> config operations could affect the RURI. Here's everything that's happening
>> until the sanity check involved:
>>
>> request_route {
>>
>>     if(is_method("KDMQ")) {
>>         dmq_handle_message();
>>     }
>>
>>     # no connect for sending replies
>>     set_reply_no_connect();
>>
>>     if($ua =~ "friendly-scanner|sipcli|sipvicious|VaxSIPUserAgent") {
>>         # silent drop for scanners - uncomment next line if want to reply
>>         # sl_send_reply("200", "OK");
>>         exit;
>>     }
>>
>>     if (!mf_process_maxfwd_header("10")) {
>>         force_rport();
>>         sl_send_reply("483","Too Many Hops");
>>         exit;
>>     }
>>
>>     # OPTIONS and NOTIFYs directed to myself
>>     if(is_method("OPTIONS|NOTIFY") && uri==myself && $rU==$null) {
>>         force_rport();
>>         sl_send_reply("200","Keepalive");
>>         exit;
>>     }
>>
>>     # All keep-alive methods regardless of destination
>>     if ( $hdr(Event) == "keep-alive") {
>>         force_rport();
>>         sl_send_reply("200","Keepalive");
>>         exit;
>>     }
>>
>>     if(!sanity_check("17895", "7")) {
>>         xlog("Malformed SIP request from $si:$sp\n");
>>         exit;
>>     }
>>
>> BR,
>> George
>>
>> On Wed, 8 Jul 2020 at 10:58, Daniel-Constantin Mierla <[email protected]>
>> wrote:
>>
>>> Hello,
>>>
>>> check your config operations, because the R-URI seems to be the next
>>> string (without quotes): "<BC><EA><8F>"
>>>
>>> You can try to print $mb in such case to see the entire SIP message
>>> buffer.
>>>
>>> Cheers,
>>> Daniel
>>> On 08.07.20 09:48, George Diamantopoulos wrote:
>>>
>>> Hello Daniel,
>>>
>>> Thanks for the reply. Indeed there is, not sure how I managed to miss
>>> that. And it wasn't about the schema after all:
>>> Jul  7 18:42:11 lbpub0-stage-lhe0-cn1 /usr/sbin/kamailio[909]: DEBUG: {1
>>> <null> 172.30.154.189 102 ACK
>>> [email protected] - sanity
>>> [sanity.c:277]: check_ruri_scheme(): check_ruri_scheme entered
>>> Jul  7 18:42:11 lbpub0-stage-lhe0-cn1 /usr/sbin/kamailio[909]: DEBUG: {1
>>> <null> 172.30.154.189 102 ACK
>>> [email protected] - <core>
>>> [core/parser/parse_uri.c:1254]: parse_uri(): uri too short: <<BC><EA><8F>>
>>> (3)
>>> Jul  7 18:42:11 lbpub0-stage-lhe0-cn1 /usr/sbin/kamailio[909]: DEBUG: {1
>>> <null> 172.30.154.189 102 ACK
>>> [email protected] - <core>
>>> [core/parser/parse_uri.c:1328]: parse_sip_msg_uri(): bad uri <<BC><EA><8F>>
>>> Jul  7 18:42:11 lbpub0-stage-lhe0-cn1 /usr/sbin/kamailio[909]: WARNING:
>>> {1 <null> 172.30.154.189 102 ACK
>>> [email protected] - sanity
>>> [sanity.c:282]: check_ruri_scheme(): failed to parse request uri
>>> [<BC><EA><8F>]
>>> Jul  7 18:42:11 lbpub0-stage-lhe0-cn1 /usr/sbin/kamailio[909]: DEBUG: {1
>>> <null> 172.30.154.189 102 ACK
>>> [email protected] - sanity
>>> [sanity_mod.c:254]: w_sanity_check(): sanity checks result: 0
>>> Jul  7 18:42:11 lbpub0-stage-lhe0-cn1 /usr/sbin/kamailio[909]: ERROR: {1
>>> <null> 172.30.154.189 102 ACK
>>> [email protected] - <script>: Malformed
>>> SIP request from 172.30.154.189:5060
>>>
>>> Still, not sure what the problem is though...
>>>
>>> BR,
>>> George
>>>
>>> On Wed, 8 Jul 2020 at 09:30, Daniel-Constantin Mierla <[email protected]>
>>> wrote:
>>>
>>>> Hello,
>>>>
>>>> when the ruri scheme check fails, there should be another debug message
>>>> saying that. Have you pasted all log messages for the failure case?
>>>>
>>>> Cheers,
>>>> Daniel
>>>> On 07.07.20 22:23, George Diamantopoulos wrote:
>>>>
>>>> Sorry, I realised I copy pasted wrong log messages for Call 1. Here's
>>>> the relevant part showing success for call 1 in contrast with Call 2:
>>>>
>>>> grep 2a859fcc4e1c8f840191a81d7c16e76d kamailio.log | egrep
>>>> 'check_ruri_scheme|w_sanity_check' | grep ACK
>>>> Jul  7 18:42:11 lbpub0-stage-lhe0-cn1 /usr/sbin/kamailio[907]: DEBUG:
>>>> {1 <null> 172.30.154.189 102 ACK
>>>> [email protected] - sanity
>>>> [sanity.c:277]: check_ruri_scheme(): check_ruri_scheme entered
>>>> Jul  7 18:42:11 lbpub0-stage-lhe0-cn1 /usr/sbin/kamailio[907]: DEBUG:
>>>> {1 <null> 172.30.154.189 102 ACK
>>>> [email protected] - sanity
>>>> [sanity.c:297]: check_ruri_scheme(): check_ruri_scheme passed
>>>> Jul  7 18:42:11 lbpub0-stage-lhe0-cn1 /usr/sbin/kamailio[907]: DEBUG:
>>>> {1 <null> 172.30.154.189 102 ACK
>>>> [email protected] - sanity
>>>> [sanity_mod.c:254]: w_sanity_check(): sanity checks result: 1
>>>>
>>>> On Tue, 7 Jul 2020 at 21:34, George Diamantopoulos <
>>>> [email protected]> wrote:
>>>>
>>>>> Hello all,
>>>>>
>>>>> I'm not 100% sure this is the only culprit in an issue I'm
>>>>> investigating, but superficially it appears that RURI scheme sanity module
>>>>> checks from the default config (flags 17895 in REQINIT) fail if the
>>>>> RURI in an ACK following a 487 includes parameters. Example from two calls
>>>>> from a kamailio instance acting as registrar/usrloc server, INVITE RURIs
>>>>> are after usrloc lookup:
>>>>>
>>>>> Call 1: INVITE sip:[email protected]:5063 SIP/2.0
>>>>> Call 2: INVITE sip:[email protected]:32768;line=moo62e08
>>>>> SIP/2.0
>>>>>
>>>>> These INVITEs produce no complaints. Later, the same registrar
>>>>> produces ACKs to acknowledge 487 (thus, same transaction ACKs) responses
>>>>> from the next proxy in the path following a CANCEL:
>>>>>
>>>>> Call 1: ACK sip:[email protected]:5063 SIP/2.0
>>>>> Call 2: ACK sip:[email protected]:32768;line=moo62e08
>>>>> SIP/2.0
>>>>>
>>>>> The next proxy (which produced/relayed the 487) processes the ACK for
>>>>> Call 1 successfully, but sanity_check at the proxy drops the request for
>>>>> Call 2 with:
>>>>>
>>>>> DEBUG: {1 <null> 172.30.154.189 102 ACK
>>>>> [email protected] - sanity
>>>>> [sanity.c:277]: check_ruri_scheme(): check_ruri_scheme entered
>>>>> DEBUG: {1 <null> 172.30.154.189 102 ACK
>>>>> [email protected] - sanity
>>>>> [sanity_mod.c:254]: w_sanity_check(): sanity checks result: 0
>>>>>
>>>>> whereas Call 1 seems OK:
>>>>>
>>>>> DEBUG: {1 <null> 172.30.154.189 102 ACK
>>>>> [email protected] - sanity
>>>>> [sanity.c:305]: check_required_headers(): check_required_headers entered
>>>>> DEBUG: {1 <null> 172.30.154.189 102 ACK
>>>>> [email protected] - sanity
>>>>> [sanity.c:313]: check_required_headers(): check_required_headers passed
>>>>>
>>>>> Could this be a bug in sanity module? Is there anything one can do in
>>>>> config which could result in illegal ACKs being produced for hop-by-hop
>>>>> transactions? schema appears to be sip: in both cases...
>>>>>
>>>>> Thank you. Best regards,
>>>>> George
>>>>>
>>>>
>>>> _______________________________________________
>>>> Kamailio (SER) - Users Mailing 
>>>> [email protected]https://lists.kamailio.org/cgi-bin/mailman/listinfo/sr-users
>>>>
>>>> --
>>>> Daniel-Constantin Mierla -- www.asipto.comwww.twitter.com/miconda -- 
>>>> www.linkedin.com/in/miconda
>>>> Funding: https://www.paypal.me/dcmierla
>>>>
>>>> --
>>> Daniel-Constantin Mierla -- www.asipto.comwww.twitter.com/miconda -- 
>>> www.linkedin.com/in/miconda
>>> Funding: https://www.paypal.me/dcmierla
>>>
>>>
_______________________________________________
Kamailio (SER) - Users Mailing List
[email protected]
https://lists.kamailio.org/cgi-bin/mailman/listinfo/sr-users

Reply via email to