[jitsi-dev] [ice4j] org.ice4j.ice.ConnectivityCheckClient.run() will skip a check beat


#1

Hey ice4j gurus,

Please excuse my ignorance but… is it OK that I see the FINEST Logger message "org.ice4j.ice.ConnectivityCheckClient.run() will skip a check beat" continuously printed seemingly forever within a call? I start a plain audio call and the message does no seem to be printed "forever". Then I turn the plain audio call into a conference audio call by adding a new participant and the message still does not get printed "forever". But as soon as I turn that into a conference video call by switching on my video (on the focus), the message starts being printed "forever".

As far as I can see in the code and if it is to be printed at all, it should happen in roughly 20 milliseconds. But within the same millisecond it often gets printed more than once. It's also common to see it printed within 5 milliseconds, not 20. And I'm already half an hour within that test conference video call… Is it OK as well?

Best regards,
Lyubomir Marinov


#2

Hi Lyubomir,

A long time has passed since you have reported this one, sorry for the delay.

This problem was related to connectivity checks which never end. It is corrected by svn revision #333 (Ice4j) and #9816 (Jitsi), but only for rfc5245 compliant ICE version (Jingle). For Gingle, it is "normal" that the connectivity checks last the whole call long.

Thank you for the report.

Best regards,
Vincent

···

On 08/02/2012 01:55 PM, Lyubomir Marinov wrote:

Hey ice4j gurus,

Please excuse my ignorance but� is it OK that I see the FINEST Logger message "org.ice4j.ice.ConnectivityCheckClient.run() will skip a check beat" continuously printed seemingly forever within a call? I start a plain audio call and the message does no seem to be printed "forever". Then I turn the plain audio call into a conference audio call by adding a new participant and the message still does not get printed "forever". But as soon as I turn that into a conference video call by switching on my video (on the focus), the message starts being printed "forever".

As far as I can see in the code and if it is to be printed at all, it should happen in roughly 20 milliseconds. But within the same millisecond it often gets printed more than once. It's also common to see it printed within 5 milliseconds, not 20. And I'm already half an hour within that test conference video call� Is it OK as well?

Best regards,
Lyubomir Marinov

--
Vincent Lucas, Ph.D. Jitsi developer
chenzo@jitsi.org http://jitsi.org