[erlang-bugs] Weird documentation of distributed applications

[PAILLEAU Eric]

Le 18/12/2012 11:54, Anthony Ramine a écrit :
> Furthermore, the documentation contradicts itself in the note following what you pasted:
> 
>> All involved nodes must have the same value for distributed and sync_nodes_timeout, or the behaviour of the system is undefined.
> 
> The list of nodes is part of the `distributed` value.
> 

Hi,

I do not agree. This sentence is true. distributed and
sync_nodes_timeout are the same on cp1, cp2 and cp3 config .
The 'problem' is only on sync_nodes_mandatory tuple.

When I told this to Loïc at Erlang Lite Paris, I was convinced that I
could not add current node to running configuration in
sync_nodes_mandatory, until I simply test again.

In the past I tested it with error when adding current node, but
this was surely only a newbie error at the time (DNS problem on FQDN
node names maybe), because
we cannot see any change in the Git repository .
But I can't remember what Erlang version I used at this time.

Maybe the erlang team may know if this part of kernel code changed
before Git integration.

Anyway the documentation might tell reader that the same configuration
is possible on all nodes.

regards.