Benjamin Lee <benjaminlee <at> users.sf.net> writes:
> > I agree with Alexander, this needs a FAQ entry. > > I first stumbled across this _behaviour_ a few years ago using a > provider M*3*6*5* which will remained unnamed. ;-P > > Since, kannel is actually behaving correctly. The solution we > implemented involved monitoring kannel status via curl and > /status?password= and detecting the links that were disconnected... > and restarting them automatically via /start-smsc?smsc= > > This solution seemed better than hacking kannel into coping with a > poor / buggy SMPP server implementation... and... the nice side affect > was that the monitoring scripts produced logs for _bad password_ / > disconnected incidents. These logs could be graphed with RRDTool -- > although I have not done this yet. > Know this is old, but just for the heck of it. I've been continuously experiencing SMPP connections just going stale, and for some aparent reason bearerbox feels this is not noteworthy to log to file (even with loglevel set to 0). Outbound queue for Africa. Restarting bearerbox or the SMSC using start-smsc?smsc= causes more pain. In both cases the connected smsbox also dies, which is a BIG problem when you are continuously sending SMSs. For a while I used tcpkill to cut the TCP connections every 15 minutes, but found out the hard way that messages are getting lost in the process. I see this problem as been around for years, with no solution is sight. <sigh>
