sequence_number is unique per bind, not per SMSC, and the typical implementation is that it is a monolithically increasing number. So if you have multiple binds to the same SMSC you won't have just a few "clashes" but almost all of them will clash :)
If via the same bind you send the same sequence_number more than once, then that is wrong, but on different binds, that's the normal way for it to work On Thu, Oct 3, 2013 at 1:18 AM, Iain Dickason <[email protected] > wrote: > Hi all, > > We have run into an issue where a carrier has been returning error code > 0x00000045 on a small number of submit_sm's. They have said that it due to > Kannel sending multiple messages with the same sequence number. > > Looking at the logs I can see that Kannel is re-using the same sequence > number for different messages on the same SMSC. An interesting fact was > that the number of duplicate sequence numbers seems to match the number of > binds we have to that SMSC. > > Reviewing the log files it shows that when the connection/SMSC isn't > highly loaded we have very few errors as the messages are handled quickly > and there is very little time for the duplicate sequence numbers to clash. > But when it is highly loaded and there is a delay in the DLR's there is a > larger time window for the sequence numbers to clash. > > The Kannel source code looks to keep the sequence number unique per bind > rather than per SMSC. Is this the expected behaviour? > > Thanks > > Iain >
