Re: [OpenSIPS-Users] Inbound DID TO/INVITE issue

2012-10-17 Thread Mike O'Connor
Hi Brett

On 17/10/12 3:52 PM, Brett Nemeroff wrote:
> Well hold on a sec..
>
> First of all, the TO field is irrelevant. So whatever RURI you have
> (that's the top line INVITE URI), that's where we're sending the call
> to next. If the below invite hits asterisk it should be delivered to
> 111610. If that's not right, you need to set your $rU to whatever you
> want it to be delivered to.
>
This is what the commercial Asterisk provider is saying.

My problem is that the sip config line
"alias_db_lookup("dbaliases","d")" changes the INVITE to our service
number and adjusts the IP address / port all the while losing the
inbound DID so I can not have one registration and a number of phone
numbers.
> Per the docs, the function you are using updates the RURI:
> http://www.opensips.org/html/docs/modules/1.7.x/alias_db.html#id250076
>
> Are you suggesting asterisk is routing on the TO header? This happens
> with some buggy SIP clients from time to time, but I wouldn't expect
> this in Asterisk.
>
No I'm saying that the TO field has the original INVITE with the DID in
it, and because Asterisk DOES NOT routing using the TO field the call is
not correctly routed. (inside the clients Asterisk PBX)
> The "To" Header really shouldn't be considered for routing. That being
> said, there are a handful of UAs out there that insist on doing so. I
> think they are pre-3261 typically but this isn't confirmed.
> -Brett
So I'm asking what should I be doing if I receive a DID INVITE, I need
to route this to a registred using with our losing the ability for the
UAs to correct route the call.

Thanks
Mike

___
Users mailing list
Users@lists.opensips.org
http://lists.opensips.org/cgi-bin/mailman/listinfo/users


Re: [OpenSIPS-Users] Inbound DID TO/INVITE issue

2012-10-17 Thread Kevin Sandy
We resolved a similar issue by having Asterisk register once for each DID. 
Alternatively, you could modify their dial plan to take the To header into 
account, but I'd go the multiple registrations route if possible. 

-- 
kevin sandy, dcap, mcp

On Oct 17, 2012, at 3:44 AM, Mike O'Connor  wrote:

> Hi Brett
> 
> On 17/10/12 3:52 PM, Brett Nemeroff wrote:
>> Well hold on a sec..
>> 
>> First of all, the TO field is irrelevant. So whatever RURI you have
>> (that's the top line INVITE URI), that's where we're sending the call
>> to next. If the below invite hits asterisk it should be delivered to
>> 111610. If that's not right, you need to set your $rU to whatever you
>> want it to be delivered to.
> This is what the commercial Asterisk provider is saying.
> 
> My problem is that the sip config line
> "alias_db_lookup("dbaliases","d")" changes the INVITE to our service
> number and adjusts the IP address / port all the while losing the
> inbound DID so I can not have one registration and a number of phone
> numbers.
>> Per the docs, the function you are using updates the RURI:
>> http://www.opensips.org/html/docs/modules/1.7.x/alias_db.html#id250076
>> 
>> Are you suggesting asterisk is routing on the TO header? This happens
>> with some buggy SIP clients from time to time, but I wouldn't expect
>> this in Asterisk.
> No I'm saying that the TO field has the original INVITE with the DID in
> it, and because Asterisk DOES NOT routing using the TO field the call is
> not correctly routed. (inside the clients Asterisk PBX)
>> The "To" Header really shouldn't be considered for routing. That being
>> said, there are a handful of UAs out there that insist on doing so. I
>> think they are pre-3261 typically but this isn't confirmed.
>> -Brett
> So I'm asking what should I be doing if I receive a DID INVITE, I need
> to route this to a registred using with our losing the ability for the
> UAs to correct route the call.
> 
> Thanks
> Mike
> 
> ___
> Users mailing list
> Users@lists.opensips.org
> http://lists.opensips.org/cgi-bin/mailman/listinfo/users

___
Users mailing list
Users@lists.opensips.org
http://lists.opensips.org/cgi-bin/mailman/listinfo/users


Re: [OpenSIPS-Users] Inbound DID TO/INVITE issue

2012-10-17 Thread Mike O'Connor
Hi Kevin

Surely there is a better solution than this, because all the Asterisk
systems I've seen have inbound routing without a registration for each DID.

The supplier of the commercially support Asterisk would need to make
changes which they are this point are not prepared to support, when
every other ITSP they have connected to in the USA and Europe, has
supported the one registration and more than one DID.

So I really need to understand what the standard should be and how my
config is incorrect.

Cheers
Mike

On 17/10/12 8:54 PM, Kevin Sandy wrote:
> We resolved a similar issue by having Asterisk register once for each DID. 
> Alternatively, you could modify their dial plan to take the To header into 
> account, but I'd go the multiple registrations route if possible. 
>


___
Users mailing list
Users@lists.opensips.org
http://lists.opensips.org/cgi-bin/mailman/listinfo/users


Re: [OpenSIPS-Users] Inbound DID TO/INVITE issue

2012-10-17 Thread Kevin Sandy
Then the other providers are doing something similar to the prior suggestion - 
replacing $rU with $tU before relaying.

One registration = one contact URI. So, multiple registrations, rewrite the 
user, or ditch the registrations and route to them by IP or host name. 


-- 
kevin sandy, dcap, mcp

On Oct 17, 2012, at 6:49 AM, Mike O'Connor  wrote:

> Hi Kevin
> 
> Surely there is a better solution than this, because all the Asterisk
> systems I've seen have inbound routing without a registration for each DID.
> 
> The supplier of the commercially support Asterisk would need to make
> changes which they are this point are not prepared to support, when
> every other ITSP they have connected to in the USA and Europe, has
> supported the one registration and more than one DID.
> 
> So I really need to understand what the standard should be and how my
> config is incorrect.
> 
> Cheers
> Mike
> 
> On 17/10/12 8:54 PM, Kevin Sandy wrote:
>> We resolved a similar issue by having Asterisk register once for each DID. 
>> Alternatively, you could modify their dial plan to take the To header into 
>> account, but I'd go the multiple registrations route if possible. 
>> 
> 
> 
> ___
> Users mailing list
> Users@lists.opensips.org
> http://lists.opensips.org/cgi-bin/mailman/listinfo/users

___
Users mailing list
Users@lists.opensips.org
http://lists.opensips.org/cgi-bin/mailman/listinfo/users


Re: [OpenSIPS-Users] Inbound DID TO/INVITE issue

2012-10-17 Thread Flavio Goncalves
Hi,

One of the best ways to solve it is to include the DID in an extra header.
We usually include something such as X-DID: did_number. Then in the
Asterisk/Freeswitch PBX server you can recover this DID reading this header
and routing internally. Many ITSPs implement in this way. Thus, you can
have a single registration but multiple DIDs.

Flavio E. Goncalves





2012/10/17 Kevin Sandy 

> Then the other providers are doing something similar to the prior
> suggestion - replacing $rU with $tU before relaying.
>
> One registration = one contact URI. So, multiple registrations, rewrite
> the user, or ditch the registrations and route to them by IP or host name.
>
>
> --
> kevin sandy, dcap, mcp
>
> On Oct 17, 2012, at 6:49 AM, Mike O'Connor  wrote:
>
> > Hi Kevin
> >
> > Surely there is a better solution than this, because all the Asterisk
> > systems I've seen have inbound routing without a registration for each
> DID.
> >
> > The supplier of the commercially support Asterisk would need to make
> > changes which they are this point are not prepared to support, when
> > every other ITSP they have connected to in the USA and Europe, has
> > supported the one registration and more than one DID.
> >
> > So I really need to understand what the standard should be and how my
> > config is incorrect.
> >
> > Cheers
> > Mike
> >
> > On 17/10/12 8:54 PM, Kevin Sandy wrote:
> >> We resolved a similar issue by having Asterisk register once for each
> DID. Alternatively, you could modify their dial plan to take the To header
> into account, but I'd go the multiple registrations route if possible.
> >>
> >
> >
> > ___
> > Users mailing list
> > Users@lists.opensips.org
> > http://lists.opensips.org/cgi-bin/mailman/listinfo/users
>
> ___
> Users mailing list
> Users@lists.opensips.org
> http://lists.opensips.org/cgi-bin/mailman/listinfo/users
>
___
Users mailing list
Users@lists.opensips.org
http://lists.opensips.org/cgi-bin/mailman/listinfo/users


Re: [OpenSIPS-Users] Error "Commands out of sync; you can't run this command now "

2012-10-17 Thread Vlad Paiu

Hello,

What version of OpenSIPS are you using ? Can you please upgrade to 1.8 
and see if the problem still appears ?


Still, this seems like a recurring issue ( already being reported by 
multiple people ), so I'll send you a patch that will help in debugging 
such issues.


Regards,

Vlad Paiu
OpenSIPS Developer
http://www.opensips-solutions.com

On 10/15/2012 02:46 AM, David Cunningham wrote:

Hello,

We recently experienced a database error "Commands out of sync; you 
can't run this command now". I understand this is a known issue with 
presence. Can anyone advise what version it was fixed in?


Thank you!

Oct 12 10:34:01 myhost /sbin/opensips[10735]: 
ERROR:db_mysql:db_mysql_submit_query: driver error on query: Commands 
out of sync; you can't run this command now
Oct 12 10:34:01 myhost /sbin/opensips[10735]: ERROR:core:db_do_query: 
error while submitting query
Oct 12 10:34:01 myhost /sbin/opensips[10735]: ERROR:auth_db:get_ha1: 
failed to query database



--
David Cunningham, Voisonics
http://voisonics.com/
USA: +1 213 221 1092
UK: +44 (0) 20 3298 1642
Australia: +61 (0) 2 8063 9019



___
Users mailing list
Users@lists.opensips.org
http://lists.opensips.org/cgi-bin/mailman/listinfo/users


___
Users mailing list
Users@lists.opensips.org
http://lists.opensips.org/cgi-bin/mailman/listinfo/users


Re: [OpenSIPS-Users] FW: Opensips 1.6.4 doesn't send an OK response under high load

2012-10-17 Thread Diego Barberio
Hi Binnan,

Thanks for all your help.

I've made a network trace and the other party is sending the TCP's ACKs
with delay, so it's seems opensips is not causing this issue.
I will keep you updated if anything else comes out.

Thanks
Diego

On Thu, Oct 11, 2012 at 6:01 PM, Binan AL Halabi wrote:

>
> Hi Diego,
> I forgot to tell that opening large number of tcp sockets means opening
> large amount of file descriptors.
>
> echo 128000 > /proc/sys/fs/inode-max
> echo 64000 > /proc/sys/fs/file-max
> ulimit -n 64000
>
> // Binan
>
>
>
>   --
> *Från:* Diego Barberio 
> *Till:* Binan AL Halabi ; OpenSIPS users
> mailling list 
> *Skickat:* fredag, 12 oktober 2012 1:38
> *Ämne:* Re: [OpenSIPS-Users] FW: Opensips 1.6.4 doesn't send an OK
> response under high load
>
> Hi Binnan,
>
> Thanks for all your help.
> I've applied all you suggestions and nothing changed.
>
> I generated a trace of OpenSIPs connections using TCP dump (attached). I
> noticed that the send queues have a lot of data between 6 pm and 8 pm (when
> is the period the problem appears).
> As far as I know this means opensips is calling send funcion but the OS is
> not actually sending data, am I right?
>
> What can be causing this problem? Do you have any other ideas?
>
> Again, thanks a lot for your help.
>
> Regards
> Diego
>
> On Tue, Oct 9, 2012 at 7:20 PM, Binan AL Halabi 
> wrote:
>
>
> Do the following to increase TCP performance:
>
> 1- Increase the number of  available local ports:
> echo 1024 65000 > /proc/sys/net/ipv4/ip_local_port_range
>
> 2- Increase the amount of memory associated with socket buffers (socket
> input and output queues):
> For input queues:
> echo 262143 > /proc/sys/net/core/rmem_max
> echo 262143 > /proc/sys/net/core/rmem_default
> And wmem_max , wmem_default for output queues.
>
> 3- TCP stack
> echo 0 > /proc/sys/net/ipv4/tcp_sack
> echo 0 > /proc/sys/net/ipv4/tcp_timestamps
>
> 4- opensips memory pool is well configured :
> http://www.opensips.org/Resources/DocsTsMem
>
>
> // Binan
>--
> *Från:* Diego Barberio 
> *Till:* OpenSIPS users mailling list 
> *Skickat:* tisdag, 9 oktober 2012 20:08
> *Ämne:* Re: [OpenSIPS-Users] FW: Opensips 1.6.4 doesn't send an OK
> response under high load
>
> Hi Binan,
>
> Thank's to Bogdan's patch I was able to run opensips 1.8.1. However, the
> problem with TCP under highload is still present.
>
> Do you have any other ideas?
>
> Thanks
> Diego
>
> On Mon, Sep 10, 2012 at 4:59 PM, Diego Barberio <
> diego.barbe...@redmondsoftware.com> wrote:
>
> Hi Binan,
>
> I tried upgrading to 1.8.1 but I'm having some issues I'm duscussin with
> Bogdan on the following thread:
>
> http://lists.opensips.org/pipermail/users/2012-August/022764.html
>
> Thanks
> Diego
>
>
> On Mon, Sep 10, 2012 at 4:47 PM, Binan AL Halabi 
> wrote:
>
> hi Diego,
>
> 1- As you dont have state in database so you dont need to check that.
> 2- Why you dont upgrade to opensips 1.8.1 ? since it contains TCP fix.
>
> Regards.
> //Binan
>
> --- On *Mon, 9/10/12, Binan AL Halabi * wrote:
>
>
> From: Binan AL Halabi 
>
> Subject: Re: [OpenSIPS-Users] FW: Opensips 1.6.4 doesn't send an OK
> response under high load
> To: "OpenSIPS users mailling list" 
> Date: Monday, September 10, 2012, 12:19 PM
>
>
> If you have
>
> --- On *Mon, 9/10/12, Diego Barberio 
> *wrote:
>
>
> From: Diego Barberio 
> Subject: Re: [OpenSIPS-Users] FW: Opensips 1.6.4 doesn't send an OK
> response under high load
> To: "OpenSIPS users mailling list" 
> Date: Monday, September 10, 2012, 10:51 AM
>
> Hi Binan,
>
> I understand what do you say with "play around TCP connection lifetime"
> however I don't get the part you say "check the database". I don't have any
> database, opensips is working without any DB.
> Which database are you talking about?
>
> Thanks
> Diego
>
> On Mon, Sep 10, 2012 at 1:37 PM, Binan AL Halabi 
> wrote:
>
> Hi Diego,
> play around TCP connection lifetime, you could find something.
>
> one thing more check the database during the busy hour and see if
> something expired.
>
> //Binan
>
>
> --- On *Mon, 9/10/12, Diego Barberio 
> *wrote:
>
>
> From: Diego Barberio 
> Subject: Re: [OpenSIPS-Users] FW: Opensips 1.6.4 doesn't send an OK
> response under high load
> To: "OpenSIPS users mailling list" 
> Date: Monday, September 10, 2012, 6:39 AM
>
>
> Hi All,
>
> I'm still having this issue. However, I went on with my investigation and
> I've discovered that this issue only happens with my SIP over TCP
> connections (I have 10 TCP connections)
> I've upgraded to version:
> Server:: OpenSIPS (1.7.2-notls (i386/linux))
>
> And the issue is still there.
>
>
> Do you have any other ideas? (Disabling TCP is not an option)
>
> Thanks
> Diego
>
>
>
>
>
> From: 
> users-boun...@lists.opensips.org
> [mailto:users-boun...@lists.opensips.org

[OpenSIPS-Users] Transfering a call by opensips

2012-10-17 Thread Engineer Voip
Hello all, 
I want to transfert the call to user C when user A calls user B in interval of 
time for example: 11h-14h
I can do that by asterisk but i prefer to do it by opensips
It's possible to do that by opensips?

Cordialement. 
Envoyé de mon iPhone
___
Users mailing list
Users@lists.opensips.org
http://lists.opensips.org/cgi-bin/mailman/listinfo/users