Looks like kamailio-debuginfo rpm was from older version of kamailio. I'm not
able to reproduce core file anymore. Could somebody please be so kind and
explain why ?
Secondly:
Mar 13 18:12:57 ricvmf-fusion01 kam-scscf[13524]: WARNING: tm
[t_lookup.c:1536]: t_unref(): WARNING: script writer di
Dan,
There are two cores because of a crash in one process followed by a
crash when the other processes are trying to shutdown.
What's interesting is that the bt doesn't show useful pointers. If you
have installed from RPMs make sure the kamailio-debuginfo is from the
same build as the other
Carsten,
Yes, this is something I saw earlier already so this attr is in my config.
Thanks for pointing this out.
On 03/13/2014 02:45 PM, Carsten Bock wrote:
It looks a little bit like a "double free".
You could try to disable the call to "abort()" in case this happens:
mem_safety=1
See: http
Jason,
I've tried multiple combinations for pattern but I'm getting only 2 core files
...
Details:
~]# cat /proc/sys/kernel/core_pattern
/tmp/core.%e.sig%s.%p
~]# lsb_release -a
LSB Version:
:base-4.0-amd64:base-4.0-noarch:core-4.0-amd64:core-4.0-noarch:graphics-4.0-amd64:graphics-4.0-noa
Hugh W. just recommended something there might be a little chance that core files
are being overwritten. Let me make them unique .. the pattern I'm using now is :
/tmp/core.%p which is not very "unique" apparently.
On 03/13/2014 02:58 PM, Jason Penton wrote:
I don't think these cores indi
I don't think these cores indicate the real crash... I'd like to get some
more detail on what actually happened? Daniel, can you re-create? Keep in
mind that if your core dump config on your box is not configured to name
your cores according to process id or timestamp one core will overwrite the
ot
It looks a little bit like a "double free".
You could try to disable the call to "abort()" in case this happens:
mem_safety=1
See: http://www.kamailio.org/wiki/cookbooks/devel/core#mem_safety
Kind regards,
Carsten
2014-03-13 19:44 GMT+01:00 Carsten Bock :
> It looks a little bit like a "double f
It looks a little bit like a "double free".
You could try to disable the call to "abort()" in case this happens:
2014-03-13 17:22 GMT+01:00 Daniel Ciprus :
> There are no more core files on the filesystem :-(
>
> On 03/13/2014 12:18 PM, Jason Penton wrote:
>
> I'm afraid this is also not the cor
There are no more core files on the filesystem :-(
On 03/13/2014 12:18 PM, Jason Penton wrote:
I'm afraid this is also not the correct core. Can you check the timestamp on
the cores? Can you re-create the crash and send me the correct core?
On Thu, Mar 13, 2014 at 5:36 PM, Daniel Ciprus
ma
I'm afraid this is also not the correct core. Can you check the timestamp
on the cores? Can you re-create the crash and send me the correct core?
On Thu, Mar 13, 2014 at 5:36 PM, Daniel Ciprus wrote:
> So I cleaned up my junkyard and I got 2 core files:
>
> (gdb) bt
> #0 0x005350b0
One note: this is not real HW, this is virtual machine ... I hope this core is
not related to some internal clock which is not in sync with HW clocks.
On 03/13/2014 11:36 AM, Daniel Ciprus wrote:
So I cleaned up my junkyard and I got 2 core files:
(gdb) bt
#0 0x005350b0 in ?? ()
#1 0
So I cleaned up my junkyard and I got 2 core files:
(gdb) bt
#0 0x005350b0 in ?? ()
#1 0x0053542a in ?? ()
#2 0x005356c7 in timer_main ()
#3 0x0046d572 in main_loop ()
#4 0x0047030b in main ()
(gdb) bt full
#0 0x005350b0 in ?? ()
No symbol tab
Hi Daniel,
this is the wrong core file. This is the one created on shutdown of
kamailio. Can you do a bt on the other core file that you probably have...
Cheers
Jason
On Thu, Mar 13, 2014 at 5:05 PM, Daniel Ciprus wrote:
> Folks,
>
> This is happening during the registration on SCSCF.
>
> Ser
Folks,
This is happening during the registration on SCSCF.
Server:: kamailio (4.2.0-dev2 (x86_64/linux))
Build:: mi_core.c compiled on 10:01:09 Mar 13 2014 with gcc 4.4.6
Flags:: STATS: Off, USE_TCP, USE_TLS, TLS_HOOKS, USE_RAW_SOCKS, DISABLE_NAGLE,
USE_MCAST, DNS_IP_HACK, SHM_MEM, SHM_MMAP, PK
14 matches
Mail list logo
