[Serusers] Alias per SIP_DOMAIN? something broken here?

Gerd Feiner g.feiner at cablesurf.de
Mon Nov 7 18:14:12 CET 2005


hi,

i run an SER which supports multiple realms.  all works fine,  
authentication, calls etc.  except one thing:

we are in the rare sitiuation of having the same number-block in two  
different cities.  i have aliases for my users (username is areacode 
+number) which is only the number (without area-code) to support  
dialing without area-code.  but in the case with the two same blocks  
in different area-codes the alias handling gets a bit confused:

# serctl alias show 88999998
<sip:0471288999998 at 04712.sip.domain.de>;q=1;expires=459311264
<sip:0471188999998 at 04711.sip.domain.de>;q=1;expires=459311264

so i have two aliases for the same number.  any of the numbers is in  
different SIP_DOMAINS, namely 04712.sip... and 04711.sip..., but even  
when i state the SIP_DOMAIN to use for the lookup stuff i get:

# SIP_DOMAIN=04712.sip.domain.de serctl alias show 88999998
<sip:0471288999998 at 04712.sip.domain.de>;q=1;expires=459311207
<sip:0471188999998 at 04711.sip.domain.de>;q=1;expires=459311207

and of course the other way round with 04711.  the realm/domain is  
stored within the alias, so there are two records:

88999998	04711.sip.domain.de	sip:0471188999998 at 04711.sip.domain.de	...	
88999998	04712.sip.domain.de	sip:0471288999998 at 04712.sip.domain.de	...

with different domains.  now, what happens is:  when a user from the  
04711 area-code calls another user (for him in that same area), he/ 
she might end up reaching a user with the 04712 area-code.  even  
though the INVITE messages and all state, that the particular (any)  
user is from the 04711 domain, or 04712 for that matter ...

i also tried to name the alias 88999998 at 04712.sip... and so on, but  
then, it doesn't even find an alias at all.

what did i miss?

oh, and yes, both realms are defined via an alias-directive in  
SER.cfg so he treats both as local (or at least it should).

any hint would be appreciated.


brgds,
g. feiner




More information about the sr-users mailing list