VoIP Mailing List Archives

[brian at freeswitch.com]

Was an issue filed in github?

On Fri, Apr 30, 2021 at 8:41 AM Kakiman Expert <jolexpert@gmail.com (jolexpert@gmail.com)> wrote:

[s.kainz at wnt.at]

Hi,
No I did not. Im still trying to find the problem, since I don’t really know if its a bug or my gross incompetence

We are testing different setups/configurations for weeks now, and we are no closer to a solution then we were a month ago …
However, we have a script now, that restarts freeswitch when sipsak can’t register anymore … which is … you know …

In the meantime we tried the following:

Setting accept-blind-reg to true ( Because we use a register.lua for authentication, and I always thought its possible I made a mistake there )
Didn’t help.

Put the database on the same host ( To cross out network-issues )
Didn’t help.

Tried different Server Hardware
Didn’t help

Tried different Freeswitch-Versions / Debian Versions
( So far Debian 10 + FS 1.10.3, Debian 8 + FS 1.4.18 ) + different Hardware Didn’t help

I will file an issue on GitHub asap, with all the information I have.

Regards,
Stefan

[gregor at infomedia.si]

I think some other user  has similar issue. Common thing is that you both are using lua for authentication. Just for info.... 

On Fri, Apr 30, 2021, 17:45 Stefan Kainz <s.kainz@wnt.at (s.kainz@wnt.at)> wrote:

[brian at freeswitch.com]

My crystal ball tells me that we can't really fix what isn't reported, we do try, but it's next to impossible to read minds.  Joking aside, Just get all you can into a github issue, we can try to sort it and solve your issue.

/b







On Fri, Apr 30, 2021 at 11:44 AM Gregor Nanger <gregor@infomedia.si (gregor@infomedia.si)> wrote:

[s.kainz at wnt.at]

To be fair, I’ve had to bother you guys a few times before on this mailing-list, and sometimes Brian, it really looks like some of you guys have a crystal ball
I have opened an issue now, with all information I can think of / remember.
https://github.com/signalwire/freeswitch/issues/1180

I will provide a backtrace as soon as this happens again and I have the chance to get one.

Thank you so much for your effort.
This issue already haunts me in my dreams xD

Thanks again and regards,
Stefan

[lists at telefaks.de]

I am having the same issue, but with registrations via XML-Curl. So this is not only LUA related. However when setting caching for registrations (e.g. cacheable="600000" msecs), so that not every registration attempt hits the XML-Curl API, this didn't change anything. In our scenario, I have the feeling that lot's of presence messages make internal profile stuck.


Am 30.04.21 um 18:25 schrieb Gregor Nanger:

[s.kainz at wnt.at]

Hallo Peter,
Seems like there are a few people having the same issue…
I also crossed out my Lua-script by setting accept-blind-reg to true, the error still persists.

Im currently testing the watchdog feature like seven1240 suggested in the my ticket on GitHub.
https://github.com/signalwire/freeswitch/issues/1180

We also have a lot of presence dialogs, but we have production-systems handling 10 times more…
So I guess that’s not it, at least it shouldn’t be.

Maybe you can also add your details to the ticket.

Regards,
Stefan

[anthony.minessale at g...]

When it's stuck you should crash it on purpose "fsctl crash" and get a "thread apply all bt" from the core in gdb.Then you might learn something.




On Mon, May 24, 2021 at 11:22 AM Stefan Kainz <s.kainz@wnt.at (s.kainz@wnt.at)> wrote: