On 05/16/06 02:48, Maxim Sobolev wrote:
> Hi,
>
> Reportedly there is an problem in radiusclient library with md5 hash
> calculation on 64bit platforms (amd64, sparc64, etc), which results in
> server rejecting requests even with proper secret configured on
> client. I have applied a fix, but don't have access to any 64bit
> hardware, therefore can't verify.
>
> Therefore, if anybody is having problems with radiusclient on 64bit
> arch he is more than welcome to try update to the latest version from
> cvs and let me know if it helps or not. Instructions on checking out
> sources from berlios cvs are here:
> https://developer.berlios.de/cvs/?group_id=1208.
>
> Thanks!
>
> -Maxim
Hi,
Reportedly there is an problem in radiusclient library with md5 hash
calculation on 64bit platforms (amd64, sparc64, etc), which results in
server rejecting requests even with proper secret configured on client.
I have applied a fix, but don't have access to any 64bit hardware,
therefore can't verify.
Therefore, if anybody is having problems with radiusclient on 64bit arch
he is more than welcome to try update to the latest version from cvs and
let me know if it helps or not. Instructions on checking out sources
from berlios cvs are here: https://developer.berlios.de/cvs/?group_id=1208.
Thanks!
-Maxim
Thanks Daniel,
But, is more quick/efficient if i use integer avp's instead of
string based avp's?
Thanks!
Ricardo.-
> -----Mensaje original-----
> De: Daniel-Constantin Mierla [mailto:daniel@voice-system.ro]
> Enviado el: Lunes, 15 de Mayo de 2006 17:14
> Para: Ricardo Martinez
> CC: 'users(a)openser.org'
> Asunto: Re: [Users] AVP usage.
>
>
> Hello,
>
> this may help you a bit. There are some examples of using
> integer-named
> avps.
> http://openser.org/docs/openser-radius-1.0.x.html
>
> Cheers,
> Daniel
>
>
> On 05/15/06 23:27, Ricardo Martinez wrote:
> > Hello.
> > I have a question. Is there a way to optimize the use of AVP's.
> > Right now i'm using avp_radius to load client's attributes
> from my database.
> > These attributes are flags with value ON/OFF. Inside my
> configuration i'm
> > using these values with avp names : callee_attr1,
> callee_attr2, etc. I
> > have read somewhere in the list that if i use numeric avp's
> instead of
> > string the processing is more quick. Is this ok?. If the
> answer is yes, how
> > is the syntax for this?.
> > For example i'm loading avp attributes like :
> >
> > SIP-AVP="attr1:on"
> > SIP-AVP="attr2:off"
> >
> > then i my configuration :
> > if ( avp_check("callee_attr1", "eq/s:on/g") ) {
> > xlog("L_INFO", "time [$Tf] : Something XXXX\n");
> > route(8);
> > return;
> > };
> >
> > Can someone guide me on this point ?
> > Thanks in advance,
> >
> > Regards,
> >
> > Ricardo Martinez.-
> >
> > _______________________________________________
> > Users mailing list
> > Users(a)openser.org
> > http://openser.org/cgi-bin/mailman/listinfo/users
> >
> >
>
Hello.
I have a question. Is there a way to optimize the use of AVP's.
Right now i'm using avp_radius to load client's attributes from my database.
These attributes are flags with value ON/OFF. Inside my configuration i'm
using these values with avp names : callee_attr1, callee_attr2, etc. I
have read somewhere in the list that if i use numeric avp's instead of
string the processing is more quick. Is this ok?. If the answer is yes, how
is the syntax for this?.
For example i'm loading avp attributes like :
SIP-AVP="attr1:on"
SIP-AVP="attr2:off"
then i my configuration :
if ( avp_check("callee_attr1", "eq/s:on/g") ) {
xlog("L_INFO", "time [$Tf] : Something XXXX\n");
route(8);
return;
};
Can someone guide me on this point ?
Thanks in advance,
Regards,
Ricardo Martinez.-
> On 05/10/2006 01:38 AM, Douglas Garstang wrote:
> > Can someone post/link to some examples of how to use avp
> module to write/read data to/from MySQL tables?
> >
> maybe the new avp_db_query() is more straightforward, but it
> is only on
> development version.
>
> http://openser.org/docs/modules/1.1.x/avpops.html#AEN268
>
> > The openser.org docs are reeeeealy bad.
> >
> Then try to improve them.
Are you kidding? How can I improve them when I don't understand the product fully, because the documentation is so bad?
I really believe that people adding to the documentation by exploring themselves is a VERY BAD idea.
When you do this, you make assumptions based on the information you currently have available to you. When that information is incomplete, you make incorrect assumptions. You find that later on, things break in certain conditions because your assumptions where bad in the first place. You also contribute to more bad information when you write documentation based on what you THINK you know, and make it available to others.
This is why the people who know the product the best, the developers, should be the ones writing the documentation.
Doug.
It is SOLVED. It took some time to realize that m_dump() act as client to
opensip server (even local) so it was firewall and opensip.cfg relay part
issue.
Thanks, Darko
Hi,
I am trying to set up message handling for offline users. Storing seems to
work ok - message is in database and all fields seems to be correct.
Dumping doesn't work - it is called like in examples - just after save
function in part handling REGISTER message but no packets at all is delivered
to registering UA (seen by ethereal), except ones need for register.
part of debug log for handling one message (11th in the database):
14(15501) DEBUG: timer routine:0,tl=0x4068410c next=(nil)
14(15501) DEBUG: final_response_handler:stop retr. and send CANCEL
(0x40683fe0)
14(15501) DEBUG:tm:t_should_relay_response: T_code=0, new_code=408
14(15501) DEBUG:tm:local_reply: branch=0, save=0, winner=0
14(15501) DEBUG:tm:local_reply: local transaction completed
14(15501) DBG: trans=0x40683fe0, callback type 256, id 0 entered
14(15501) MSILO:m_tm_callback: completed with status 408 [mid: 1080573260/11]
14(15501) MSILO:m_tm_callback: message <11> was not sent successfully
14(15501) MSILO: msg_list_set_flag: mid:11 fl:8
14(15501) DEBUG: add_to_tail_of_timer[2]: 0x40684028
14(15501) DEBUG: final_response_handler : done
TIA, Darko
I thought that was already done? I remember asking this several weeks back, and was told that t_replicate() would support multiple destinations in the next release.
> -----Original Message-----
> From: Bogdan-Andrei Iancu [mailto:bogdan@voice-system.ro]
> Sent: Wednesday, May 10, 2006 1:25 PM
> To: Douglas Garstang
> Cc: Edson; unplug; users(a)openser.org
> Subject: Re: [Users] clustering
>
>
> yes, that's right - I will take a look to see how difficult
> is to change
> t_replicate() to be able to sent to more destinations in parallel.
>
> regards,
> bogdan
>
> Douglas Garstang wrote:
>
> >Then, if system N fails, all systems >N also fail. That's
> not very redundant.
> >
> >
> >
> >>-----Original Message-----
> >>From: Edson [mailto:4lists@gmail.com]
> >>Sent: Wednesday, May 10, 2006 9:01 AM
> >>To: Douglas Garstang; 'Bogdan-Andrei Iancu'; 'unplug'
> >>Cc: users(a)openser.org
> >>Subject: RE: [Users] clustering
> >>
> >>
> >>Create a 'chain' of t_replicate (1->2->3->...->n). Than You
> >>can have 'n'
> >>Proxies in the 'cluster'...
> >>
> >>Edson.
> >>
> >>
> >>
> >>>-----Original Message-----
> >>>From: users-bounces(a)openser.org
> >>>
> >>>
> >>[mailto:users-bounces@openser.org] On
> >>
> >>
> >>>Behalf Of Douglas Garstang
> >>>Sent: quarta-feira, 10 de maio de 2006 11:49
> >>>To: Bogdan-Andrei Iancu; unplug
> >>>Cc: users(a)openser.org
> >>>Subject: RE: [Users] clustering
> >>>
> >>>Yes, but t_replicate() only allows you to replicate to ONE
> >>>
> >>>
> >>other OpenSER
> >>
> >>
> >>>system. Two OpenSER systems is not a redundant configuration - 3 is
> >>>redundant.
> >>>
> >>>
> >>>
> >>>>-----Original Message-----
> >>>>From: Bogdan-Andrei Iancu [mailto:bogdan@voice-system.ro]
> >>>>Sent: Wednesday, May 10, 2006 5:28 AM
> >>>>To: unplug
> >>>>Cc: users(a)openser.org
> >>>>Subject: Re: [Users] clustering
> >>>>
> >>>>
> >>>>Hi,
> >>>>
> >>>>try yo google on the mailing list - the topic was often
> >>>>debated. search
> >>>>also for t_replicate() and replication.
> >>>>
> >>>>regards,
> >>>>bogdan
> >>>>
> >>>>unplug wrote:
> >>>>
> >>>>
> >>>>
> >>>>>How to do the clustering of openser? Any reference?
> >>>>>
> >>>>>_______________________________________________
> >>>>>Users mailing list
> >>>>>Users(a)openser.org
> >>>>>http://openser.org/cgi-bin/mailman/listinfo/users
> >>>>>
> >>>>>
> >>>>>
> >>>>>
> >>>>_______________________________________________
> >>>>Users mailing list
> >>>>Users(a)openser.org
> >>>>http://openser.org/cgi-bin/mailman/listinfo/users
> >>>>
> >>>>
> >>>>
> >>>_______________________________________________
> >>>Users mailing list
> >>>Users(a)openser.org
> >>>http://openser.org/cgi-bin/mailman/listinfo/users
> >>>
> >>>
> >
> >
> >
>
>
Hi,
in this case you will to use an algorithm that uses some information
that is common to the original call and the new call. If you are not
able to link the calls (logically) you will not be able to send them to
same destination.
regards,
bogdan
Douglas Garstang wrote:
>We're using algorithm 0, which hashes over the callid, but that doesn't help.
>The polycom phones send a new invite with a NEW callid when you hit transfer.
>
>Doug.
>
>
>
>>-----Original Message-----
>>From: Bogdan-Andrei Iancu [mailto:bogdan@voice-system.ro]
>>Sent: Wednesday, May 10, 2006 4:01 PM
>>To: Douglas Garstang
>>Cc: users(a)openser.org
>>Subject: Re: [Users] Stateful Transfers
>>
>>
>>Hi,
>>
>>what dispatching algorithm are you using?
>>
>>regards,
>>bogdan
>>
>>Douglas Garstang wrote:
>>
>>
>>
>>>I got a big problem here that I don't know how to fix.
>>>
>>>An OpenSER system forwards calls from phones to Asterisk
>>>
>>>
>>boxes. Up until now we had been using the dispatcher to
>>evenly distribute calls to the Asterisk boxes. However, when
>>someone transfers a call, using the dispatcher, the call can
>>go to a different Asterisk box, which breaks the transfer.
>>
>>
>>>I need to find some way to make OpenSER route transfers to
>>>
>>>
>>the same Asterisk system. I'm not sure how to do this. If
>>there was some way to, upon an INVITE, check and see if there
>>was already a call in progress for that user and then send it
>>to the same Asterisk system. Would stateful processing help
>>in this way? Can stateful processing 'remember' that there's
>>a call already in progress like this, or does the stateful
>>simply relate to stateful processing of a single call from
>>INVITE to OK/ACK?
>>
>>
>>>Doug.
>>>
>>>
>>>
>>>_______________________________________________
>>>Users mailing list
>>>Users(a)openser.org
>>>http://openser.org/cgi-bin/mailman/listinfo/users
>>>
>>>
>>>
>>>
>>>
>
>
>
