Re: Kamailio change rtpengine in mid-cal

Myng via sr-users <[email protected]> Wed, 29 Apr 2026 14:42:13 +0200
Newsgroups gmane.comp.voip.ser
Message-ID <CAGBhiAn5iY2v5mPaeB5MRN9uXxp4uW4UoA+2gqoLOo0GFpWiOg@mail.gmail.com>
Hi Sergey,
Thanks for your answer.

Just want to clarify, my rtpengines work independently (no redis or share
state or HA configured). I also do not expect a failover since I did not
make any specific configuration for it.





On Wed, Apr 29, 2026 at 7:52 AM Sergey Safarov <[email protected]> wrote:

> I see you make rtpengine node failover.
> I have not tested failover + reinvite.
> I think current behaviour need to be documented on rtpengine github ticket.
>
> Sergey
>
> On Wed, Apr 29, 2026 at 8:49 AM Sergey Safarov <[email protected]>
> wrote:
>
>> Please check
>> https://github.com/sipwise/rtpengine/issues/1241
>>
>> It looks like this works as expected on
>> dbf79cb76274b182f4fb1be10ea894d46b9ba818
>> <https://github.com/dvillaume/rtpengine/commit/dbf79cb76274b182f4fb1be10ea894d46b9ba818>
>> commit.
>> You can try make git bisect to find commit where this functionally is
>> broke.
>> Then you can report the ticket and wait when it will be fixed.
>>
>> Sergey
>>
>> On Wed, Apr 29, 2026 at 8:39 AM Myng via sr-users <
>> [email protected]> wrote:
>>
>>> I am encountering an issue regarding how Kamailio handles rtpengine
>>> node failures mid-call, and I would like to clarify the expected behavior.
>>>
>>> *The Scenario:* I have a setup with multiple rtpengines (e.g., 4 nodes)
>>> all configured in the same pool with setid=0.
>>>
>>>    1.
>>>
>>>    An initial INVITE arrives, and Kamailio successfully selects
>>>    *rtpengine-A* for the call.
>>>    2.
>>>
>>>    Later in the dialog, a re-INVITE occurs.
>>>    3.
>>>
>>>    Coincidentally, *rtpengine-A* is unavailable at that exact moment.
>>>    4.
>>>
>>>    Kamailio then automatically switches to a different rtpengine node
>>>    in the set.
>>>
>>> *The Question:* Based on my reading of an older issue (#2713
>>> <https://github.com/kamailio/kamailio/issues/2713>), my understanding
>>> was that Kamailio should *not* change the rtpengine node mid-call, as a
>>> new node will not have the media state for the ongoing session.
>>>
>>> Could you please confirm if my understanding is correct, or if
>>> Kamailio’s default failover behavior for mid-call transactions has changed
>>> recently?
>>>
>>>
>>> Thank you for your time and insights!
>>>
>>>
>>> __________________________________________________________
>>> Kamailio - Users Mailing List - Non Commercial Discussions --
>>> [email protected]
>>> To unsubscribe send an email to [email protected]
>>> Important: keep the mailing list in the recipients, do not reply only to
>>> the sender!
>>>
>>

__________________________________________________________
Kamailio - Users Mailing List - Non Commercial Discussions -- [email protected]
To unsubscribe send an email to [email protected]
Important: keep the mailing list in the recipients, do not reply only to the sender!