Skip to main content
A server is a listener your customers bind in to. Adding one is a port, a name, and then the policy you want applied to everyone who connects to it.

​
Adding one

Servers → Add server. Name and port are the only required fields; everything else has a working default.
Every setting is at Listener settings.
A listener that cannot take its port does not stop the gateway. It logs and everything else keeps running — right for a multi-listener deployment, and surprising the first time.A clean start is not proof every listener is up. Check /ops/health.

​
What a bind is checked against, in order

1

The credential

systemId and password must match an enabled SMPP credential. See credentials.yml.
2

allowedIps, if set

Empty means no restriction.
3

Connection limits

srv.maxConnections overall, srv.maxConnectionsPerIP, and conf.maxConnectionsPerUser.default per account.
4

The product, if conf.product.required is on

The resolved product must name an enabled row in the product catalogue. Blank or unknown is refused with 0x40F — at bind, not at submit, because the product lives on the session context.

​
Where the product comes from

system_type on the bind request, falling back to the credential’s product field.
system_type is free text validated nowhere else. A customer binding prem instead of premium passes any non-empty test and lands in the unrated state where messages are delivered and charged nothing.That is why conf.product.required checks membership of the catalogue rather than presence of a value — and why, in file mode, where there is no catalogue, only the non-empty half applies. It is logged once per configuration load naming the listener.

​
What a listener applies to everyone bound to it

​
The states a listener can be in

The panel shows these as one badge per card. Listening with nobody bound is not a fault — a listener with no customers connected is quiet, not broken, which is why it is not painted red the way a vendor with no session is.

​
Dropping one customer’s bind

Session ids come from fireflo health. They are the listener’s own counter, so they collide between listeners and restart with the gateway.
This ends a connection; it does not withdraw permission. A healthy client rebinds within seconds unless you also disable its login.

​
Inbound messages

mo.owner.map decides which account owns which inbound address. mo.owner.default is where an unmatched one goes — and empty means nowhere: the listener logs message.mo.unmapped and consumes it rather than guessing an owner. Deliberate, because delivering a stranger’s message to the wrong customer is worse. But unmapped traffic disappears quietly unless you watch for that line.