I am using dispatcher module to spread the load across multipe gateways. Is
it possible to force OpenSER reload the dispatcher.list via a fifo command?
I can't seem to find one.
Hello to all
my Openser is running, but not accepting any connections....
I did ngrep, and the client keeps sending the registry message, but
openser doenst reply.
With this configuration, openser is listening for clients in port 5060
or 5061 ?
debug=3
fork=yes
log_stderror=yes
check_via=no
dns=yes # (cmd. line: -r)
rev_dns=no # (cmd. line: -R)
port=5060
children=4
fifo="/tmp/openser_fifo"
disable_tls = 0
listen = tls:XXX.XXX.XXX.XXX:5061
tls_verify = 1
tls_require_certificate = 1
tls_method = TLSv1
tls_certificate ="/tls/tools/rootCA/cacert.pem"
tls_private_key ="/tls/tools/rootCA/private/cakey.pem"
tls_ca_list ="/tls/tools/rootCA/certs/01.pem"
Thanks
Joao Pereira
Hi,
May be loosely related: you could check ongoing discussion on
sip(a)ietf.org mailing list, thread "draft-ietf-sip-outbound".
They are discussing outbound-proxy and discovery concepts. Sometimes
they are touching similar problem that you've described.
--
Regards,
Arek Bekiersz
Klaus Darilion wrote:
> Hi!
>
> This is a little bit OT, but maybe you can help me out. Consider the
> following setup: I have a SIP service running on mydomain.com.
>
> My sip AoR is sip:klaus@mydomain.com
>
> mydomain.com resolves to my main SIP proxy (registrar, proxy, used by
> all local clients) using NAPTRs, SRV and A records.
>
> Further, I do want to have a separate proxy (edge proxy) which handles
> signaling from/to external domains, TLS, white/blacklisting ....
>
> Thus, the domain mydomain.com should resolve to:
> - the main proxy for SIP clients
> - the edge proxy for SIP proxies
>
> Of course I could use different domains or the clients have to configure
> an outbound proxy (which points directly to the main proxy), but this
> complicates configuration.
>
> AFAIK, Microsoft solved this problem within the LCS using different
> prefixes during the SRV lookup. The Live Communicator (client) uses
> _sipinternal._udp.mydomain.com to find the main proxy, whereas the LCS
> uses _sip._udp.domain.com to find the edge proxy (access proxy in
> Microsoft terms) of mydomain.com.
>
> Thus, I would like to raise discussion: Have you an idea how to solve
> this? Are you aware of any RFCs/drafts addressing this problem?
>
> regards
> klaus
Hi!
This is a little bit OT, but maybe you can help me out. Consider the
following setup: I have a SIP service running on mydomain.com.
My sip AoR is sip:klaus@mydomain.com
mydomain.com resolves to my main SIP proxy (registrar, proxy, used by
all local clients) using NAPTRs, SRV and A records.
Further, I do want to have a separate proxy (edge proxy) which handles
signaling from/to external domains, TLS, white/blacklisting ....
Thus, the domain mydomain.com should resolve to:
- the main proxy for SIP clients
- the edge proxy for SIP proxies
Of course I could use different domains or the clients have to configure
an outbound proxy (which points directly to the main proxy), but this
complicates configuration.
AFAIK, Microsoft solved this problem within the LCS using different
prefixes during the SRV lookup. The Live Communicator (client) uses
_sipinternal._udp.mydomain.com to find the main proxy, whereas the LCS
uses _sip._udp.domain.com to find the edge proxy (access proxy in
Microsoft terms) of mydomain.com.
Thus, I would like to raise discussion: Have you an idea how to solve
this? Are you aware of any RFCs/drafts addressing this problem?
regards
klaus
Hello openser experts,
I was trying to set up a openser box in public IP and observed a
wiered behaviour.
If I ran openser in port 5060 and try to register from a sip
client, I get response from the openser, but when I try to run it in
different port, something like 40000. and then try to connect from a sip
client, I could see that the packets could reach the openser box using
ethereal. But openser does not respond to these requests. I tried to put log
in route() but didnt get any, this looks like openser didnot pick up the
request for itself.
I am using openser 1.0.1 and I am not running any firewall in the
public openser.
Is this problem known ? Am I missing something ?
Also what is the significance of "listen" parameter in openser
configuration file.
Regards,
Voipwala
Hello,
I am using openser with mediaproxy.
When mediaproxy is used for a call it is activated in two steps:
- on the caller side at the reception of the invite
- on the callee side at the reception of the 200 ok
In fact when the sdp body is present in these messages.
However what happens when the callee does not answer quicky enough.
Because the mediaproxy already handles the communication with the caller and no audio is exchanged yet
the idle timeout expires (default 60 seconds) and when the callee answers the call the audio does not work.
Is this behavior correct ?
Best regards
Jean-Luc Behr
Hi everyone,
I have a question about the dispatcher and rr-module:
I thought about the following setup:
OpenSER with the dispatcher on one machine which does loadbalancing (and
maybe later Failover) and several SIP-Proxies behind the dispatcher. The
SIP-Proxies should be easily exchangeable so the loose-routing should be
done via the dispatcher-Server. I have the following setup in mind
(simplyfied, in my juvenile carelessness):
Dispatcher (IP:62.153.141.6)
route{
# Round robin: Every Request to another Server.
ds_select_dst("1", "4");
forward(uri:host, uri:port);
}
SIP-Proxy:
route {
##################################################################################################################
# Loose-Routing
###############################################################################################################
if (loose_route()) {
if (!t_relay()) {
log(1, "Not possible to relay\n");
# Fehler melden
sl_reply_error();
return;
}
}; # if (loose_route()) {
##################################################################################################################
# Record-Route
###############################################################################################################
record_route_preset("62.153.141.6:5060");
}
Unfortunately this Setup does not work like i thought... :-(
The messages seem to be looping between Dispatcher and SIP-Proxy. Any hints what i could do or how i could solve this dilemma? Has anyone (i guess so) tried anything like this before? Can anyone point me to working example configs?
I need to do the record_route_preset() on the selected SIP-Proxy, because i set a user-specific From-Line with the UAC-Module, which stores it's data in the Record-Route-Header.
I need to do record routing for accounting and i do not want to do accounting on the dispatcher...
Thanks in advance,
Carsten
Hello to all
after some problems to put openser to work with TLS and RTPproxy, the
server is now running... but shows this messages:
WARNING: rtpp_test: of RTP proxy <unix:/var/run/rtpproxy.sock>doesn't
support required protocol version 200503224(9656)
WARNING: rtpp_test: support for RTP proxy
<unix:/var/run/rtpproxy.sock>has been disabled temporarily
did anyone had this problem?
Thanks
Joao Pereira
Hi,
Few days ago I presented UA that sends public IP in "Contact" and "Via"
HF, when it is behind NAT. It is in private LAN but it learns its public
IP and every SIP request will contain public addresses.
It ruins NAT traversal effort, inluding nathelper + rtpproxy. I cannot
properly detect NAT-ed user neither by "Via" nor "Contact" check. As a
result people with various modems get angry and are reporting to me all
those "I have no audio" problems. Or they cannot even register at all,
if they have partcularly stupid NAT.
What is your advice?
Vendor knows about issue but doesn't see problem.
Of course I can enable NAT traversal by "User-Agent" HF check, but 90%
of those UAs are NOT behind NAT...
--
Regards,
Arek Bekiersz