Got an update on this from the firewall's perspective. Apparently the
Barracuda was in (SIP) Proxy mode. After turning that off and just using
plain NAT everything worked like expected.
Cheers,
Dirk
___
SIP Express Router (SER) and Kamailio (OpenSER) -
This was also my understanding. But as you can understand I want to be
certain. Currently this issue only rises with the Cisco hardware in
combination with the Barracuda firewall. Other appliances seem to
respect the order properly.
On 07-10-16 09:30, Daniel-Constantin Mierla wrote:
> I think the
Hi,
Looking at an issue with SIP-Cisco and a Barracuda firewall, I've come
across something I need to know more about.
The setup is as follows:
CLIENT -> KAMAILIO -> ASTERISK
CLIENT <- KAMAILIO <- ASTERISK
Client sends INVITE, Kamailio forwards to Asterisk, SIP 100 and SIP 183
are being sent and
Hi Daniel,
Did you manage to get a chance to look at all the dumps I sent?
Cheers,
Dirk
On 18-07-16 11:58, Daniel-Constantin Mierla wrote:
> If you didn't get other crashes, I can't think of anything else that can
> help from the backtrace.
>
> Cheers,
> Daniel
>
On 07/18/2016 11:58 AM, Daniel-Constantin Mierla wrote:
>
> OK -- just to have full picture -- did you experienced more crashes than
> the two ones you reported?
Yes, we experienced three more crashes
I will send you these gdb bt full crashes directly.
> I see -- can you describe a bit the chan
On 07/18/2016 10:49 AM, Daniel-Constantin Mierla wrote:
> have you run with -x qm before reverting to 4.2?
No we didn't, as I didn't know what the effects would be, it's a live
server so we don't want to try changes before testing.
> Is the same config you run with 4.2 and 4.4?
Not at the moment,
heers,
> Daniel
>
>
> On 15/07/16 13:06, Dirk Teurlings - Signet B.V. wrote:
>> (gdb) frame 1
>> #1 dlg_unref (dlg=dlg@entry=0x7f585c494b40, cnt=cnt@entry=1) at
>> dlg_hash.c:921
>> 921 dlg_lock( d_table, d_entry);
>> (gdb) p *dlg
>> $1
-Constantin Mierla wrote:
> From the second crash, can you get:
>
> frame 1
>
> p *dlg
>
> So far it looks like either to a double free or some buffer overflow...
>
> Cheers,
> Daniel
>
>
> On 15/07/16 10:51, Dirk Teurlings - Signet B.V. wrote:
>>
Hi, first crash:
(gdb) info locals
cb =
__FUNCTION__ = "run_dlg_callbacks"
(gdb) bt full
#0 run_dlg_callbacks (type=type@entry=64, dlg=dlg@entry=0x7fceb400e2f0,
req=req@entry=0x7fced4f093c8, rpl=rpl@entry=0x0, dir=,
dlg_data=dlg_data@entry=0x0)
at dlg_cb.c:253
cb =
__FUNCTIO
ld process 12376 exited by a signal 11
Jul 15 10:42:26 server /usr/sbin/kamailio[12370]: ALERT:
[main.c:742]: handle_sigs(): core was generated
Jul 15 10:42:26 server /usr/sbin/kamailio[12370]: INFO:
[main.c:754]: handle_sigs(): terminating due to SIGCHLD
Cheers,
Dirk
On 07/15/2016 10:06 AM,
Hi,
Running Kamailio on Debian from the Kamailio repository with 4.4.2
stable (unpatched). Getting some random segfaults with it now, here's
the relevant backtrace from the generated core.
Core was generated by `/usr/sbin/kamailio -f /etc/kamailio/kamailio.cfg
-P /var/run/kamailio/kamailio.'.
Pro
On 07/06/2016 03:53 PM, Daniel-Constantin Mierla wrote:
> Hello,
>
> can you try with latest branch 4.4 or master? I just pushed a commit for it.
Yes, database is now found and entry is removed as expected. The normale
event_route[dialog:end] is however almost never fired still.
But I can work ar
Hi,
There's something going on with a locally originated BYE and the dialog
module running with a database backend. Here's the scenario.
Call gets setup, and is being monitored by outside processes
(non-Kamailio). After a while these processes determine the call is
taking too long (or too expensi
On 06/27/2016 11:01 AM, Daniel-Constantin Mierla wrote:
> as far as I can remember, sending a redirect 3xx response takes in
> consideration any new request URI, which is added as contact. Now, not
> sure why the second parameter of get_contacts() would influence that, no
> much time to look at the
On 06/22/2016 07:45 AM, Daniel-Constantin Mierla wrote:
> Looking quickly at the code, both versions of the functions are
> executing the same inner function, but there can still be some
> conditions on the reason parameter, so I need to dig in further.
>
> However, I need to understand first what
Hi,
Testing with latest stable 4.4 Kamailio at the moment, and running into
the following issue.
I have a setup that sends a 302 to do a Moved Temporarily to another
extension. To update the Contact header accordingly, I use the
get_redirects function in the failure_route.
get_redirects('*', 'Re
Hi,
I'm seeing a difference in timestamp for acc modules recorderd INVITES
and the dialog start event.
This is the log entry:
Apr 1 12:55:59 host /usr/sbin/kamailio[15642]: INFO:
you send sources.list which contains Kamailio Debian Stable Repos?
>
> Regards.
>
>> On 29 Jan 2016, at 16:40, Dirk Teurlings - SIGNET B.V.
>> wrote:
>>
>> Hi,
>>
>> While compiling Kamailio 4.3.4 from package sources for Debian 7 Wheezy
>> (ka
Hi,
While compiling Kamailio 4.3.4 from package sources for Debian 7 Wheezy
(kamailio repository) I'm running into a problem. I've forked the CNXCC
module which I was trying to keep up to date with upstream changes. But
before I could do that, there is a dependency of redis added.
Unfortunately, i
On 25-01-16 12:49, Daniel-Constantin Mierla wrote:
> It should not matter the version, there was no change to is_myself() in
> the recent releases
Thanks everyone, works like a charm!
Cheers,
Dirk
___
SIP Express Router (SER) and Kamailio (OpenSER
On 22-01-16 11:31, Daniel-Constantin Mierla wrote:
> try with is_myself("sip:$fU@fd").
Hi Daniel,
Thanks for the quick response. Comparing the from_uri to myself should
only compare IP or domain name correct? So wouldn't this be enough as a
domain check?
is_myself("$fd")
Cheers,
Dirk
__
Hi,
The documentation states the following.
https://www.kamailio.org/wiki/cookbooks/4.2.x/core#myself
if(uri==myself) {
log("the request is for local processing\n");
};
Now this could also be used to do something like this: (also used in
some examples)
if(from_uri==myself) {
log("the reques
I'm seeing the exact same message sometimes as well.
For me it seems to be tied to volume, as soon as a lot of calls are
coming from an interconnect this could occur. Also, as soon as this
happens both incoming and outgoing calls have audio problems.
My RTP instance isn't as much patched as y
On 30-09-15 13:29, Fred Posner wrote:
Without a version of rtpproxy using the -A flag, you'll need to either
(1) update to a different version of rtpproxy or (2) skip rtpproxy and
have your asterisk handle all the rtp.
I tried rtpproxy v2, with the -A flag in bridge mode ( -A
privateip/public
On 30-09-15 12:23, Fred Posner wrote:
Are you using -A flag in rttproxy?
Unfortunately we're running a version of RTPPROXY at the moment that
doesn't have this flag. I'm considering upgrading to rtproxy 2.0, but
will need to test whether this doesn't affect anything else.
It's preferred t
On 30-09-15 11:28, Daniel Tryba wrote:
A quick and dirty workaround might be the topoh module (which might break
other things though).
Already tried that as well, but that doesn't mask the private IP in the
important places. It only hides them in the via and record-route headers.
Cheers,
Dir
On 30-09-15 11:10, Daniel Tryba wrote:
The asterisk doesn't need to know the network topology beyond its connection
with kamailio, so you need to fix it at the place of traversal between public
and private: kamailio.
Agreed, I was looking for a workaround at this point, but to no avail.
Cheer
Hi,
I found some existing topics on this but failed to get a solutions out
of them.
We're running into some issues with client devices connecting to our
private addresses. The way it is setup now:
CLIENTS <-> (NAT) <-> INTERNET <-> KAMAILIO(4.2.5) with
RTPPROXY(v1) <-> PRIVATE LAN <->
Although we also use CNXCC I can't confirm this bug, mainly because we
only fire up CNXCC from routes when the user still has credits
(controlled and checked by MySQL using transactions)
As far as the lock_get goes, look into the cnxcc_check.c to find out
what gets released when the user overf
d only for requests
received from the network, not for those generated locally.
Cheers,
Daniel
On 21/07/15 10:05, Dirk Teurlings - SIGNET B.V. wrote:
Hi again,
Running kamailio 4.2.5 with the acc module to do accounting in a MySQL
database. We are now running into the issue that sometimes there is n
Hi again,
Running kamailio 4.2.5 with the acc module to do accounting in a MySQL
database. We are now running into the issue that sometimes there is no
BYE line when there is an INVITE.
Digging into this I noticed that the dialog:end event does get
triggered, and that CNXCC consumes the call
ssion.
I guess for further insight into your issue, you would have to sahre
your routing-logic.
Thanks,
Carsten
2015-07-14 12:15 GMT+02:00 Dirk Teurlings - SIGNET B.V. :
Hello,
Running kamailio 4.2.5 for a few weeks now, coming from 4.2.3. Since the
upgrade we increased the logging subst
Hello,
Running kamailio 4.2.5 for a few weeks now, coming from 4.2.3. Since the
upgrade we increased the logging substantially in our routing to get
more insight into what is going on.
We noticed that in a few cases the log showed a SIP message was being
threated by multiple forks, each cons
Hi,
$var is in a scope per thread, so you will have a different counter in
every thread.
You should use shared memory instead:
http://www.kamailio.org/wiki/cookbooks/4.2.x/pseudovariables#shv_name_-_shared_memory_variables
Cheers,
Dirk
On 19-06-15 15:54, Ali Taher wrote:
Hello,
I want t
ialog and upper layer modules relying on it. I
will push a patch for it.
Cheers,
Daniel
On 16/06/15 13:59, Dirk Teurlings - SIGNET B.V. wrote:
Hello,
Following up on this issue, though KA seems to work fine now. We're
running into another problem. We use a fork of CNXCC that terminates
all
Hello,
Following up on this issue, though KA seems to work fine now. We're
running into another problem. We use a fork of CNXCC that terminates all
calls when the user doesn't have enough credit. But with these patches
that doesn't work anymore and the current thread even locks up completely.
now I understand that the dialog is actually not ended, right?
Cheers,
Daniel
On 07/05/15 17:05, Dirk Teurlings - SIGNET B.V. wrote:
Hi Daniel,
Good to see you found an improvement, not sure yet if it'll help us.
Here is the non-debug log first. Just now I noticed that the init
Hi Daniel,
Good to see you found an improvement, not sure yet if it'll help us.
Here is the non-debug log first. Just now I noticed that the initial
dialog doesn't get ended as well now, 5fa4dc3670596fb626269f083b0c9480
in the log.
The setup is like this
Leg 1 Kamailio -> Asterisk
L
d 4? It will save me
some time to look into the code and fix the issue.
Cheers,
Daniel
On 07/05/15 08:46, Dirk Teurlings - SIGNET B.V. wrote:
Hi,
(Kamailio 4.2.3)
Working to get dialogs closed and cleaned up nicely I'm looking into
the ka_interval and ka_timer parameter of the dialog module
Hi,
(Kamailio 4.2.3)
Working to get dialogs closed and cleaned up nicely I'm looking into the
ka_interval and ka_timer parameter of the dialog module. The following
is now happening.
Scenario:
1. Call gets setup
2. Callee's internet connection drops
3. Call gets terminated (nicely) due to ka
M
To: Dirk Teurlings - SIGNET B.V.
Subject: Re: [SR-Users] Transaction good practices with t_relay t_newtran
and t_release
Dirk,
Why do you need to create new transactions yourself with t_newtran()?
Nonexotic applications of Kamailio usually don't require this. A transaction
is automaticall
Hi,
In our config we don't use presence, but I do have a question about
transactions.
We're using dialogs as well as transactions in the routing. To keep
things simplyfied, let's assume we have the following stripped out route
setup:
request_route {
if (is_method("CANCEL")) {
On 17-03-15 07:30, Daniel-Constantin Mierla wrote:
use:
kamcmd pkg.stats
Excellent! exactly what we were looking for.
Cheers,
Dirk
___
SIP Express Router (SER) and Kamailio (OpenSER) - sr-users mailing list
sr-users@lists.sip-router.org
http://lis
Hi,
We'd like to be able to get statistics of the used amount of memory in
Kamailio.
# kamctl stats shmem
The above command only tells us what the shared memory is doing. Is
there any way to get the committed size of memory for each of the
PKG_MEMORY pools?
If not, could this be added at
44 matches
Mail list logo
