Re: [mqtt-comment] Server redirection

["Ken Borgendale" <kwb@us.ibm.com>]

We are currently outside of a public comment period.

The MQTTv5.0 specification does not define how server redirection works other than that the server is allowed to pass a string property (Server Reference) in CONNACK and DISCONNECT when using the 0x9C(Use another server) or 0x9D (Server moved) return codes. These return codes and the property can be used by the client but the client is not required to do so. The spec does not define the format of the string, nor the processing at either server or client
This is a compromise as there was interest in the function but not the desire to fully specify the behavior of it.

As the client is by definition disconnected at the point of the redirect, presumably the server does the same thing it does for any other case of the client being disconnected.


Ken Borgendale -- kwb@us.ibm.com

<mqtt-comment@lists.oasis-open.org> wrote on 12/05/2017 10:20:10 AM:

> From: David Katz <David.Katz@bmw-carit.de>
> To: "mqtt-comment@lists.oasis-open.org" <mqtt-comment@lists.oasis-open.org>
> Date: 12/05/2017 10:25 AM
> Subject: [mqtt-comment] Server redirection
> Sent by: <mqtt-comment@lists.oasis-open.org>
>
> Hi all,
>  
> 4.11 Server redirection
>  
> Question: Given a persisted session and subscriptions >= 1, what provisions are made for
> publications made to a topic to which the client had subscribed after the client
> redirects? It should not be allowed to redirect a client for which there are
> publications waiting. Also, if it is to be allowed to redirect a client with persisted
> sessions, the race conditions must be considered: what happens if a publication arrives
> at the moment that the redirect is sent?
>  
> Cheers,
> David Katz
>  
> BMW Car IT GmbH
> David Katz
> Moosacher Straße 86
> 80809 München

>