Hello, Henning, I am worried about this scenario, because it's a symptom of what may happen in other cases. For instance, I've seen that this operator usually sends re-invites immediate after sending ACK. This may create race conditions like 3.1.5 of RFC5407 https://tools.ietf.org/html/rfc5407#page-22 I'd understand that one happens because of packet loss, as it's in UDP's nature, but in this case it would be artificially created by Kamailio. if there was no problem at network level (packet loss, packets following different path on the network and arriving out of order), why Kamailio creates it? I'd expect that the shared memory is used precisely for this. If an instance of kamailio receives a 200 OK, it could check on the shm and say "hey, another instance is processing a 180 for this call. Let's wait for it to finish" (*). I know there could still be a problem, the instance processing the 180 undergoes a context switch just after it receives the message, but before writing to shm, but it would greatly reduce the chance. In our applications we use a SIP stack that always sends messages to the application in the same order it receives them, even though is multi-threaded and messages from the network are received by different threads. So, they really syncronize between them. Why Kamailio instances don't? I am evaluating kamailio to use it as a dispatcher to balance load against our several Application Servers, to present to the operator just a couple of entrance points to our platform (they don't want to establish connections to each one of our servers). This operator is very difficult to deal with. I am sure they will complain something like "why are you sending messages out of order? Fix that". The operator will be able to see traces and check that messages entered the Kamailio nodes in order and left out of order. They will not accept it. (*) Not really "wait", as it would introduce a delay in processing all messages. it should be like putting it on a queue, continue processing other messages, and go back to the queue later. Well, thanks for your answer. Luis On 4/8/20 3:01 AM, Henning Westerholt wrote: Hello Luis, as the 1xx responses are usually send unreliable (unless you use PRACK), you should not make any assumption on the order or even the arrival of this messages. It can also happens on a network level, if send by UDP. Can you elaborate why you think this re-ordering is a problem for you? One idea to enforce some ordering would be to use the dialog module in combination with reply routes and the textops(x) module. About the shared memory question – Kamailio implement its own memory manager (private memory and shared memory pool). Cheers, Henning -- Henning Westerholt – https://skalatan.de/blog/ <https://nam02.safelinks.protection.outlook.com/?url=https%3A%2F%2Fskalatan.de%2Fblog%2F&data=02%7C01%7C%7C9909a729fd8a426f81aa08d7db8aab0a%7Cab4a33c2b5614f798601bc921698ad08%7C0%7C1%7C637219260993836600&sdata=ZLmPqvbWKbsXY49s870sElN2I0uIn0DtDQSqJOoxr6I%3D&reserved=0> Kamailio services – https://gilawa.com <https://nam02.safelinks.protection.outlook.com/?url=https%3A%2F%2Fgilawa.com%2F&data=02%7C01%7C%7C9909a729fd8a426f81aa08d7db8aab0a%7Cab4a33c2b5614f798601bc921698ad08%7C0%7C1%7C637219260993836600&sdata=Hdgzfwgu80wiwJBOjh9N70hvXSvWjt8abuKFjVRsavo%3D&reserved=0> *From:* sr-users &lt;sr-users-bounces(a)lists.kamailio.org&gt; &lt;sr-users-bounces(a)lists.kamailio.org&gt; *On Behalf Of *Luis Rojas G. *Sent:* Tuesday, April 7, 2020 10:43 PM *To:* sr-users(a)lists.kamailio.org *Subject:* [SR-Users] Kamailio propagates 180 and 200 OK OUT OF ORDER Good day, I am testing the dispatcher module, using Kamailio as stateless proxy. I have a pool of UAC (scripts in SIPP) and a pool of UAS (also scripts in SIPP) for the destinations. Kamailio version is kamailio-5.3.3-4.1.x86_64. Problem I have is, if UAS responds 180 and 200 OK to Invite immediately, sometimes they are propagated out of order. 200 OK before 180, like this : UAS is 172.30.4.195:5061. UAC is 172.30.4.195:5080. Kamailio is 192.168.253.4:5070 Difference between 180 and 200 is just about 50 microseconds. My guess is that both messages are received by different instances of Kamailio, and then because of context switches, even though the 180 is received before, that process ends after the processing of 200. However, I had the idea that in order to avoid these problems the kamailio processes synchronized with each other using a shared memory. I tried using stateful proxy and I obtained the same result. By the way, anyone has any idea about how Kamailio's share memory is implemented? It clearly does not use the typical system calls shmget(), shmat(), because they are not shown by ipcs command. Before posting here I googled, but I couldn't find anything related to this. I can't believe I am the only one who ever had this problem, so I guess I am doing something wrong... Please, any help. I'm really stuck on this. Thanks. -- -- Luis Rojas Software Architect Sixbell Los Leones 1200 Providencia Santiago, Chile Phone: (+56-2) 22001288mailto:luis.rojas@sixbell.com <luis.rojas@sixbell.com>http://www.sixbell.com _______________________________________________ Kamailio (SER) - Users Mailing List sr-users(a)lists.kamailio.org https://lists.kamailio.org/cgi-bin/mailman/listinfo/sr-users

If you think about it, if the 200 OK is so close to the 180 it doesn’t really matter from a signalling standpoint if the 180 comes first or if it arrives after the 200 OK. It’s the 200 OK that is important. If the 180 comes after, it’s simply ignored and the dialog is established successfully. The 1xx is seldom significant (unless you have PRACK but that’s another story). Or do you really have a situation where the 180 is critical? /O

On Thu, Apr 09, 2020 at 08:14:02AM -0400, Luis Rojas G. wrote: Agreed that you need to conserve signalling messages end-to-end, as a matter of principle; can't just drop things. As I mentioned earlier, we had the same problem--ironically, with a major mobile operator--sending reinvites and e2e ACKs almost contemporaneously. We solved it with 'async' by delaying all reinvites by 50 ms, and haven't had a single complaint since. Aside from that specific scenario, we haven't seen ordering problems, and have never had cause to call into question whether we can 'rely' on Kamailio to conserve ordering. There are valid questions raised in this thread about whether any user-space SIP element subject to the vicissitudes and realities of packet-switched networking can be relied upon to preserve ordering in a consistent and universal way. At the end of the day, this is a network problem. PSTN interworking is imperfect; it imposes synchronous assumptions upon asynchronous media, and we see this play out in many areas, e.g. fax. Not much you can do about it, but thankfully the corner cases are relatively few. -- Alex

Hi Luis, Rather confusingly, there is an 'async_workers' parameter in the core as well, which needs to be set: https://www.kamailio.org/wiki/cookbooks/5.3.x/core#async_workers There is some relationship between 'async_workers' in the core and the 'workers' modparam in the 'async' module which is explained by Daniel somewhere in a past mailing list thread, but I do not remember it offhand. I really should dig it out and update the documentation with this nuance. Having said that, I did not use the 'async' framework for my fix, but rather 'mqueue' and 'rtimer'. I have no real justification for that; just custom, habit and comfort with those mechanisms. I grew accustomed to them at a time when I had some issues with 100% CPU utilisation in a virtualised (Xen) environment when using early versions of the 'async' concepts. 1) The first thing is to create a reinvite queue; a single one will do, since it's specifically designed to be multiprocess-safe: loadmodule "mqueue" modparam("mqueue", "mqueue", "name=reinvite_q"); 2) Then I create 12 'rtimer' processes to consume this queue, each having a 10,000 usec re-invocation delay: loadmodule "rtimer" modparam("rtimer", "timer", "name=reinvite_q1;interval=10000u;mode=1;") modparam("rtimer", "exec", "timer=reinvite_q1;route=REINVITE_DEQUEUE") modparam("rtimer", "timer", "name=reinvite_q2;interval=10000u;mode=1;") modparam("rtimer", "exec", "timer=reinvite_q2;route=REINVITE_DEQUEUE") modparam("rtimer", "timer", "name=reinvite_q3;interval=10000u;mode=1;") modparam("rtimer", "exec", "timer=reinvite_q3;route=REINVITE_DEQUEUE") modparam("rtimer", "timer", "name=reinvite_q4;interval=10000u;mode=1;") modparam("rtimer", "exec", "timer=reinvite_q4;route=REINVITE_DEQUEUE") modparam("rtimer", "timer", "name=reinvite_q5;interval=10000u;mode=1;") modparam("rtimer", "exec", "timer=reinvite_q5;route=REINVITE_DEQUEUE") modparam("rtimer", "timer", "name=reinvite_q6;interval=10000u;mode=1;") modparam("rtimer", "exec", "timer=reinvite_q6;route=REINVITE_DEQUEUE") modparam("rtimer", "timer", "name=reinvite_q7;interval=10000u;mode=1;") modparam("rtimer", "exec", "timer=reinvite_q7;route=REINVITE_DEQUEUE") modparam("rtimer", "timer", "name=reinvite_q8;interval=10000u;mode=1;") modparam("rtimer", "exec", "timer=reinvite_q8;route=REINVITE_DEQUEUE") modparam("rtimer", "timer", "name=reinvite_q9;interval=10000u;mode=1;") modparam("rtimer", "exec", "timer=reinvite_q9;route=REINVITE_DEQUEUE") modparam("rtimer", "timer", "name=reinvite_q10;interval=10000u;mode=1;") modparam("rtimer", "exec", "timer=reinvite_q10;route=REINVITE_DEQUEUE") modparam("rtimer", "timer", "name=reinvite_q11;interval=10000u;mode=1;") modparam("rtimer", "exec", "timer=reinvite_q11;route=REINVITE_DEQUEUE") modparam("rtimer", "timer", "name=reinvite_q12;interval=10000u;mode=1;") modparam("rtimer", "exec", "timer=reinvite_q12;route=REINVITE_DEQUEUE") 3) I then add this handling for reinvites to the loose_route() block in the main request route -- logging and other extraneous matter omitted: --- if(has_totag()) { if(loose_route()) { if(is_method("INVITE")) { if(!t_suspend()) { sl_send_reply("500", "Internal Server Error"); exit; } mq_add("reinvite_q", "$T(id_index):$T(id_label)", ""); } else { # Normal in-dialog request handling, t_relay() and that. route(IN_DLG_REQ); } } } --- And the handler on the other side, when the transaction is reanimated, as it were: --- route[REINVITE_DEQUEUE] { while(mq_fetch("reinvite_q")) { xlog("L_INFO", "[R-REINVITE-DEQUEUE:$ci] -> Resuming re-invite handling ($TV(Sn)) in PID $pp\n"); $var(id) = $(mqk(reinvite_q){s.select,0,:}{s.int}); $var(label) = $(mqk(reinvite_q){s.select,1,:}{s.int}); xlog("L_INFO", "[R-REINVITE-DEQUEUE:$ci] -> Resuming re-invite handling ($TV(Sn)) in PID $pp\n"); # Call route[IN_DLG_REQ] to do what we otherwise would have done # immediately, were it any other kind of in-dialog request. t_continue("$var(id)", "$var(label)", "IN_DLG_REQ"); return; } } --- The idea behind 12 whole 'rtimer' processes is to massively overprovision the amount of reinvite handlers available relative to the actual number of reinvites passing through the system, even at high volumes. The danger with too few processes is that the intended 10 ms delay may not apply, because while(mq_fetch(...)) will just spin, always dequeueing new reinvites and never allowing the execution route to return control to 'rtimer'. As a practical matter, the system is not extremely high-volume and this is a very academic concern. Overall, I would describe this method as crude but effective. It was developed some years ago. 'async' provides some simplification and removes some of the manual labour around this kind of management, from what I understand, and overall I think you'd find it easier to use that. ¡Suerte! I'd be happy to answer any additional questions. -- Alex -- Alex Balashov | Principal | Evariste Systems LLC Tel: +1-706-510-6800 / +1-800-250-5920 (toll-free) Web: http://www.evaristesys.com/, http://www.csrpswitch.com/ _______________________________________________ Kamailio (SER) - Users Mailing List sr-users(a)lists.kamailio.org https://lists.kamailio.org/cgi-bin/mailman/listinfo/sr-users

Thanks for all the details. Wow, such nice implementation. It requires a deep knowledge of Kamailio.

I see, this is the part where the transaction is saved.     mq_add("reinvite_q", "$T(id_index):$T(id_label)", ""); I didn't know much about pseudo-variables. I am reading now. https://www.kamailio.org/wiki/cookbooks/5.3.x/pseudovariables#tmx_module_ps…

I guess, this approach could help me also in case of replies, not just requests, to add a delay to 200 OK to Invite.

By the way, I assigned a value to "async_workers", and now the Invite (I was delaying any Invite, not just reinvite, just to test the api) is propagated delayed, which is great, but at the same time a SIP/2.0 500 I'm terribly sorry, server error occurred (1/TM)