The issue seems to have subsided, I can't get it to crash anymore. I'm
guessing maybe something in the environment that changed. For the heck
of it I also tried the static musl libc version on the other server that
has not had issues, and it has been running for > 1 week just fine.
I'll post back with some debugging if it ever happens again.
Lance
On 4/6/2014 8:53 AM, Guus Sliepen wrote:
On Tue, Apr 01, 2014 at 01:08:56PM -0600, Lance Fredrickson wrote:
On a recently formatted debiann 7 server, I'm running tinc1.1pre9 and
running in to this type of error every 3-4 days.
Mar 31 19:35:26 MediaBox kernel: [256033.433285] tincd[3301] general
protection ip:406961 sp:7fff9e763b20 error:0 in tincd[400000+28000]
I initially had compiled tinc1.1pre9 creating a static binary using musl
libc. Thinking it must have been an issue with musl I recompiled against
the system libraries, but am still hitting this issue. Another older debiann
7 system has run for months without issue, so I'm not sure why this server
is having issues.
Was the older system also running the same static binary that was compiled
using musl? In any case, could you recompile it with debugging flags (-g -O0),
make sure you enable coredumps (ulimit -c unlimited), and run it in the
foreground (-D)? This way, when it crashes it should leave a core dump that can
be read by a gdb. Could you then either send me the tincd binary and the core
file, or send me the backtrace (gdb ./tincd core -ex 'bt full')?
Also, what kind of machine is MediaBox?
_______________________________________________
tinc mailing list
[email protected]
http://www.tinc-vpn.org/cgi-bin/mailman/listinfo/tinc
_______________________________________________
tinc mailing list
[email protected]
http://www.tinc-vpn.org/cgi-bin/mailman/listinfo/tinc
