[Standards] Solving the not staying connected in a MUC problem

Mickaël Rémond mremond at process-one.net
Thu Jun 20 07:30:38 UTC 2019


In ejabberd we are somewhat doing the reverse. We notify the users in the room when the service is shutting down to let them know they are no longer in the room.

The rationale is that we do not know when the server will be back up, and that the state may be incorrect for quite some time.
This also makes sense in the case of ejabberd as the upgrade of some components can be made without restarting the server. Many deployments are also in cluster, so shutting down a node will have part of the chat rooms migrated on the remaining nodes.

That said we miss a way to notify the users that the service is back up and that they could join again. Having a standard way to do this would be handy.


Mickaël Rémond

> On 19 Jun 2019, at 11:26, Daniel Gultsch <daniel at gultsch.de> wrote:
> Thank you Kim for explaining that. I wasn’t aware that Prosody was doing this.
> FWIW I’m totally fine with that solution as well (if Server developers
> don’t consider that too resource heavy).
> I’m in no way attached to the proposal I made. I just want to move to
> a solution that doesn’t involve my mobile client pinging 20 different
> mucs every 60 seconds.
> cheers
> Daniel
> _______________________________________________
> Standards mailing list
> Info: https://mail.jabber.org/mailman/listinfo/standards
> Unsubscribe: Standards-unsubscribe at xmpp.org
> _______________________________________________

More information about the Standards mailing list