[OpenSIPS-Devel] [ opensips-Bugs-3292984 ] CRITICAL:core:ip_addr2a: unknown address family 0

[SourceForge.net]

Bugs item #3292984, was opened at 2011-04-26 08:30
Message generated for change (Comment added) made by 
You can respond by visiting: 
https://sourceforge.net/tracker/?func=detail&atid=1086410&aid=3292984&group_id=232389

Please note that this message will contain a full copy of the comment thread,
including the initial issue submission, for this request,
not just the latest update.
Category: None
>Group: trunk
Status: Open
>Resolution: Accepted
>Priority: 3
Private: No
Submitted By: https://www.google.com/accounts ()
Assigned to: Anca Vamanu (anca_vamanu)
Summary: CRITICAL:core:ip_addr2a: unknown address family 0

Initial Comment:
Hi Bogdan,

This issue seems to be related to the following bug:
https://sourceforge.net/tracker/?func=detail&atid=1086410&aid=3164711&group_id=232389

I'm doing the B2B top hiding scenario. All my modules are built based on one of the latest trunks --- for sure they are newer than the time when you add the "Fixed" status to the previous bug report :)

In my B2B proxy I have:

local_route {
        $var(enable_nat) = 1;
        $var(debug_level) = 3;

        if ($var(debug_level) > 2) { xlog("[leg2][$si]=========LOCAL_ROUTE ($rm - $rr) - CallID=$ci=========\n"); }
        if (is_method("INVITE")) {
                fetch_dlg_value("testdlg","$avp(s:call_setup)");
                cache_store("memcached_local","b2b_call_info_$ci","$avp(s:call_setup)",14400);

                exit;
        }
        else if (is_method("ACK") && $ci=~"B2B\.[0-9]+\.[0-9]+") {
                # start billing
                route(19);
        }
        else if (is_method("BYE") && $ci=~"B2B\.[0-9]+\.[0-9]+" && src_ip=='1.1.1.1') {
                # end billing
                route(91);
        }
}


After running over 300 calls, suddenly my B2B instance died last night... and there were blood everywhere :)
I have added is_method("BYE") few days ago. Before that, since I put it in production 1 month ago, the B2B was running just fine.


Here is the log I found (B2B = leg2):

Apr 26 05:58:46 ineeddids os[16371]: [leg1](INVITE - <null>)----- ROUTE[1] -> t_relay SUCCESS
Apr 26 05:58:46 ineeddids os[16414]: [leg2][1.1.1.1]=========LOCAL_ROUTE (INVITE - <null>) - CallID=B2B.615.8054538=========
Apr 26 05:58:55 ineeddids os[16419]: CRITICAL:core:ip_addr2a: unknown address family 0
Apr 26 05:58:55 ineeddids kernel: [2291485.324537] opensips[16419]: segfault at 0 ip 00007f32c0d75e52 sp 00007fff0da362e8 error 4 in libc-2.12.1.so[7f32c0cf3000+17a000]



Please advise.

Thanks
-- Kamen

----------------------------------------------------------------------

>Comment By: https://www.google.com/accounts ()
Date: 2011-04-27 21:58

Message:
Hi Anca,

Thank you once again for your prompt response.

Tracing the calls deeply, brought me to very strange behavior from the B2B
instance. For example, I have been seeing ACKs like \0x00\0x00...... :)

I have removed the dialog,nathelper and usrloc modules and implemented
what I need with the other available options and tools.

I can clearly say the noticed issues are now gone.

I will play with it a bit longer and let you know if something crash again
:)

Thanks
-- Kamen

----------------------------------------------------------------------

Comment By: Anca Vamanu (anca_vamanu)
Date: 2011-04-27 10:58

Message:
Hi Kamen,

The obvious problem is that you are using dialog with B2B - it does not
work right ( I have now written this in the documentation also ). But let's
see exactly where this crash occurs. If you get a core file, can you please
investigate it with gdb and print here the backtrace?

Regards,
Anca

----------------------------------------------------------------------

You can respond by visiting: 
https://sourceforge.net/tracker/?func=detail&atid=1086410&aid=3292984&group_id=232389