[Dxspider-support] Multiconnects (Luigi)

[Dirk Koopman]

Luigi Carlotto, IK5ZUK wrote:
> Hi Saul,
> thak you for your reply. Reading the Changes file I understand the 
> maxconnect_user variable has the function to allow an incoming users to 
> have (as default) up to 2 other connections to other nodes. This is 
> applied only on incoming connections, and should work by checking the 
> maximum number of times a user is already connected to the rest of the 
> cluster cloud.
> 
> BTW, some days ago I've made a test on some nodes managed by myself and 
> running DX Spider Cluster version 1.55 (build 0.32) on Linux. On that 
> nodes I've done set/var $main::maxconnect_user = 1 and after that I've 
> connected with the same call and without different SSIDs (of course!) at 
> the same time without any problem.
> So, due to this, I have a doubt: or I'm missing something, or I do not 
> understand the idea of the maxconnect feature, or this new feature is 
> not working properly.
> 
> I would like to read something from Dirk...
> 

Sorry about the delay, I have rather busy on other things lately.

I have just tested it again and it *is* working for me. However, it may 
be that something else is interacting when one is testing this feature.

Firstly: it only works on incoming connections to your node.

Secondly: it only works if the PC92 for any other connections have 
arrived at your node, that is if stat/userc <call> (or sh/route <call>) 
shows that the user already has the maximum number of nodes allowed.

One of the things that may not be obvious is that it can take up to 60 
seconds before an PC92^A or ^D update for a user is generated. This is 
reduce the effect of users that use programs to login, emit a spot, then 
log out have on the network. No PC92 updates will normally be generated 
for such behaviour.

Dirk