Thank you Nikos. We don't have the flexibility of using multiple accounts so we'll just have to do the concatenation on the backend then.
Gerald 2010/9/7 Nikos Balkanas <nbalka...@gmail.com>: > No. There are SMScs out there that will recognise sockets (or IPs) and will > send each part of a concatenated SMS to the same socket. Some others do (as > is in your case). The only thing you can do that works, is to use different > accounts to connect from each kannel. > > BR, > Nikos > ----- Original Message ----- From: "Bopolissimus Platypus Jr" > <bopolissimus.li...@gmail.com> > To: <users@kannel.org> > Sent: Tuesday, September 07, 2010 5:44 AM > Subject: concatenated MOs whose parts arrive at different kannels? > > > Hello, > > We have a setup where we have two separate kannel instances each > binding (RX and TX) to the same two SMSCs. > > Only RX Connections shown below since TX is not relevant. > > Kannel1 <--- SMSC1 > | > <--- SMSC2 > > > Kannel2 <--- SMSC1 > | > <--- SMSC2 > > In previous testing with just one Kannel, different parts of a > concatenated SMS can arrive via both SMSCs (setting smsc-id the same > for logically equivalent SMSCs allows concatenation to work correctly > in this case). > > This almost certainly extends to having two kannels with this setup > (conjecture, not tested since that's a production system).. The parts > of a concatenated SMS might arrive via 1 each of the four connections > so Kannel1 might end up with parts 1 and 3 and Kannel2 might end up > with parts 2 and 4. > > We have not previously needed to use sms-combine-concatenated-mo since > there was previously no concatenated SMS support for this service. > However, we are now looking into implementing concatenated message > support and therefore are looking into options for automatic SMS > combination in the above setup. > > Apart from turning off sms-combine-concatenated-mo and recombining the > parts at the backend, are there any options for supporting combining > at the kannels (somehow)? There was a conjecture among my colleagues > that sqlbox might be a possibility. However, after looking at the > sqlbox userguide and some of the code, I don't see anything like that. > > Thanks for any pointers, or for showing where we might be going wrogn. > > Gerald > > -- > Gerald Timothy Quimpo http://bopolissimus.blogspot.com > bopolissimus.li...@gmail.com bopolissi...@gmail.com > > Even Tom Lane said: "Or, if you're worried > about actions from functions, use a trigger > to do the logging. There are approximately > no cases where a rule is really better than > a trigger :-( " > > -- Gerald Timothy Quimpo http://bopolissimus.blogspot.com bopolissimus.li...@gmail.com bopolissi...@gmail.com Even Tom Lane said: "Or, if you're worried about actions from functions, use a trigger to do the logging. There are approximately no cases where a rule is really better than a trigger :-( "
