[ Due to muted/frozen track from JVB the user become to Inactive status ]

Hi there!

Some times ago we have started to face the issue when during the conference some of the participants change their status to “inactive” without any explicit reasons.

Could you please clarify in which cases the user can catch the “inactive” status?

As I see in the code its related to the muted/frozen track from JVB, actually in:

ParticipantConnectionStatusHandler._getNewStateForJvbMode()

But I have no idea why JVB decide to change the track to become muted.

It would be useful to have some understanding of it to improve our infrastructure to avoid such situations in future.

Thank you in advance.

1 Like

Inactive when the track is out of lastN. Do you have lastN enabled, have you checked the received messages, when this happen it should be out of lastN?

Thank you for your time!

Currently I have lastN: -1 and the situation is pretty bad.

This morning I had a meeting with two colleagues and I saw when inactive status was from one to the other participant for a few minutes.
But at the same time, they both saw me and each other perfectly.

ohh… I have a ton of logs))
I have a screenshot from the console when such a situation is happened(in attachment).

one of them says that: “is in last N: false” - do you mean it?
BTW: I just set lastN as Jisti production is configured. Hope it will help somehow.

Is this meet.jit.si or your own deployment?

It’s my own deployment based on the latest stable version.

Do you see inactive icon (grey gsm bars) or ninja icon?

It’s ninja icon appears for few seconds and after video backs again. And situation repeats in random time. Few times per hour or few times per minute

This sounds like bandwidth problems …

Good day!

Could you please clarify on which side this bandwidth problem? JVB or Client?
If it’s client-side which exactly side? my bandwidth or remote participant bandwidth? (in case when I see the ninja icons on remote participants)

By the way, my JVBs are located in Amazon AWS. And I am located in an office that has two ISP providers with a summary bandwidth of more than 150Mbit/s. Also neet to notice that problem appears recently.

Might be you have some advice on how I can check the situation to find where exactly the problem is?

It can be both … You see ninja icons as jvb decided you don’t have enough bandwidth and is switching off videos. It can be client bandwidth, but it can be jvb one, but then that will happen to most of the participants.

Good day!
I would like to add a few more info about this problem. I have solved the issue only after totally disabling the bandwidth calculation on the JVB side.

org.jitsi.videobridge.TRUST_BWE=false

I think it can be configured on both client and JVB sides. Would you mind telling me how it can be configured in a less aggressive way? Maybe some thresholds or some timeouts which I can configure?

@Alex_Soloviov what’s your experience been like with BWE disabled? It’s a concept I’ve been mulling…

There is definitely fixed my problem. The “Ninja” icon does not appear anymore without explicit reasons.
But I’m sure it can be configured and it can be possible to use a such useful feature.

Good to know. Thanks for sharing!