Hi Stipe,

Bug reported [1] - spent some time tracing it, so far have got about as
far as bb_smscconn.c: smsc2_rout and smscconn.c: smscconn_usable. I've
been through the smsbox to see the smsc_id set when the messages are
split but I can't find any routing based on smsc_id at all.

>From what I can tell smsc2_rout simply starts at a random place in the
list of smsc's and checks to see if each is preferred or ok to us. The
first ok is selected unless a preferred one is found. Sound about right?

I have retested (reordering my smsc in conf to ensure the list is built
in a different order) only to find that any attempt to set the smsc_id
is ignored. How do the smsc cgi to set smsc_id on outbound work at all? 

ie: with http response including header X-Kannel-SMSC: TEST2

Smsbox log with some additional debugging ...

2003-11-28 12:02:21 [5] DEBUG: sms_split: orig smsc <TEST2>, part smsc
<TEST2>
2003-11-28 12:02:21 [5] DEBUG: message length 14, sending 1 message

Bearerboxlog with additional debugging ...

DEBUG: smscconn_usable: checking to see if conn is usable for msg - smsc
id <TEST1>
DEBUG: smsc2_rout: best_ok <TEST1>
DEBUG: smscconn_usable: checking to see if conn is usable for msg - smsc
id <MODEL>
DEBUG: smscconn_usable: checking to see if conn is usable for msg - smsc
id <TEST2>

Note my config doesn't include any allowed or preferred sections in smsc
configurations (except for MODEL connection). So it seems to run only
random (ie: cgi variables are ignored). 

To fix this could we add a smsc2_find call in the smsc2_rout (maybe
prior to the random search) thus the smsc_id is always set to
originator. But then of course all messages would get routed back
through the smsc they came from - not what we want. Could the the
smsc_id be set NULL if unforced - so no initial find would be triggered?

Cheers,
Alan

[1] http://bugs.kannel.org/view_bug_page.php?f_id=0000101

On Thu, 2003-11-27 at 13:22, Stipe Tolj wrote:
> Alan McNatty wrote:
> > 
> > Just a question really. I've been looking at the X-Kannel headers and
> > noted that if you receive an MO message from SMSC 'A' and you specify
> > the message should go back via SMSC 'A' for normal short message fine.
> > However if you respond with content with content that gets split. The
> > first fragment goes through the correct SMSC ('A') but the other
> > fragments get sent in normal load-balancing fashion.
> > 
> > Should all content go through the specified SMSC or is the current
> > implementation by design?
> 
> yep. All split message parts should go throuigh that specific smsc-id.
> If they don't then this can be considered as a bug.
> 
> Please add a bug report to http://bugs.kannel.org/ and include some
> bearerbox/smsbox log lines to show the behaviour. The better you
> describe, the faster we get it fixed. Thanks in advance.
> 
> Stipe
> 
> mailto:[EMAIL PROTECTED]
> -------------------------------------------------------------------
> Wapme Systems AG
> 
> M�nsterstr. 248
> 40470 D�sseldorf, NRW, Germany
> 
> phone: +49.211.74845.0
> fax: +49.211.74845.299
> 
> mailto:[EMAIL PROTECTED]
> http://www.wapme-systems.de/
> -------------------------------------------------------------------
> 
> -----BEGIN PGP PUBLIC KEY BLOCK-----
> Version: GnuPG v1.2.2 (Cygwin)
> 
> mIsEP6mcYwEEAMDnUiUwrbb+xwTFWN6TxF2+XZu7/alwJMeCwMBRvXtPZqfjpPhS
> OkBpU0F4TrVuugz1HINTSaJTYq10AzDQXp5NkyWgckqW79nPAWuOX0dicbJk+cN2
> nM2TI4KaxUDe6u8hghNEnH/i2lXsUu9apnP/iixzV81VC2je3uc9hZpnAAYptEVT
> dGlwZSBUb2xqIChUZWNobm9sb2d5IENlbnRlciAmIFJlc2VhcmNoIExhYikgPHRv
> bGpAd2FwbWUtc3lzdGVtcy5kZT6ItAQTAQIAHgUCP6mcYwIbAwYLCQgHAwIDFQID
> AxYCAQIeAQIXgAAKCRABV0w1BqPYRuSqA/wPzsQxao2YePENCtgRTrO86U6zg3sl
> OcS6CJFI4FZP5h/xD3GRsNH1+MPSvZlomDdpFnr547DGz/Kq9MXuQwVvlVig5yWZ
> K5dtKp1r5YLhxJQBhfirZbRFFnYmf19f18J8OoS28tuFVftDl1AIwJS3HLyBTv6H
> g2HyLAEKQIp30Q==
> =aYCI
> -----END PGP PUBLIC KEY BLOCK-----


Reply via email to