Hi,
I have a client (broken) that, for yet unknown reason, sent 2 http GET requests to REST API, which caused Janus to not respond with keepalive to any of those requests and eventually destroy that session.
Is this expected behaviour ?
Version: 1.2.4 tag
Thanks
What do you mean by “not respond with keepalive”? The GET is used as a way to perform long polling and receive responses/events from Janus: while that GET can act as a keepalive trigger (it’s activity on the session), it’s not its main task. There’s a specific keep-alive message that can be sent too (and which other transports use).
I’m using freeswitch-mod-janus which does only http GET and if there is no activity on session in 30sec a keepalive message is sent by Janus and received.
Practically built with this in mind
The long-poll request has a 30 seconds timeout. If it has no event to report, a simple keep-alive message will be triggered:
But in this case, when two consecutive http GETs are sent at same second, there is no keepalive produced by janus in next 2min (which is basically my session + reclaim_session timeout)
I tried that locally and it works as expected for me.
Edit: it was not the same second for me, though, the second GET happened maybe a second later.
Thanks for testing.
I guess it can be when it is exactly same moment sent out, but nevertheless, client shouldn’t do that 