Hi,

I assume, you are using 1.2.1 stable kannel version?! please try with 1.3.2
devel version that is _more_ stable and many many bugs were fixed...

Alejandro Guerrieri wrote:

> Hi,
> 
> I've finally found a way dealing with multiple SMSC connections to the
> same operator.
> 
> I've found it in an <almost> forgotten thread from somelist.com I've got
> from the Google cache (considering somelist.com is down since some time...
> I love Goggle cache!).
> 
> What I did is to set two SMSC connections with different names, and I've
> set preferred-smsc-id for the first connection to itself.
> 
> I also put all other connections on the denied-smsc-id list.
> 
> I send the messages using sms-push interface and setting "&smsc=balance1".
> If for some reason the link for balance1 is down, it automatically routes
> using balance2.
> 
> --
> group = smsc
> smsc = smpp
> smsc-id = balance1
> preferred-smsc-id = balance1
> denied-smsc-id = "other1;other2;another"
> host = "my.host1.address"
> port = 1234
> receive-port = 1234
> smsc-username = "somename"
> smsc-password = "somepass"
> 
> group = smsc
> smsc = smpp
> smsc-id = balance2
> denied-smsc-id = "other1;other2;another"
> host = "my.host2.address"
> port = 1234
> receive-port = 1234
> smsc-username = "somename"
> smsc-password = "somepass"
> --
> 
> It would seem like my problems are solved? NO :( keep reading:
> 
> 1. For some unknown reason, I cannot change the reconnect-delay timer. It
> keeps asking each 10 seconds despite I've set 60 seconds (as my carrier
> asked me to do) to no avail. I even tocuhed
> 
> 2. The worst problem: Kannel does not detect when the link goes down at
> SMPP level. I mean: For example, if my provider changes the password for a
> transmitter, the link goes down  and I start seeing the bind_receiver
> messages again, but the link still appears as "connected" using the http
> interface. Worst: Kannel still try to route messages thru that link, so
> all my fancy setups are close to useless since Kannel still tries to use
> the (now broken) link.
> 
> Any hints?
> 
> Thank you in advance,

-- 
Thanks,
Alex


Reply via email to