g few lines of
logs after computer crash.
Also maybe gwlist_search() and octstr_compare() could be optimized
further? :)
wapbox/bearerbox is from the latest cvs.
Mindaugas
# opreport --long-filenames | head -20
CPU: AMD64 processors, speed 2390.33 MHz (estimated)
Counted CPU_CLK_UNHALTED e
> against losing few lines of
> logs after computer crash.
So no ideas about fflush()? And how to debug "stuck" Kannel from my
previous
post? I could even give access to offending system.
Mindaugas
er - can I safely comment out fflush() after vfprintf? Or
have
configure or configuration file setting for that? :)
As I've already said - I do not care if I'll loose few lines of log in a
power outage.
Mindaugas
. Sorry but debug level log
is
not possible - log quite quickly fills all available disk space.
Mindaugas
n level servers Centos
4.x is more
common than FC or Debian. :)
Mindaugas
particular reason not to increase this to e.g. 1300 and
get max download size ~330kb?
Thanks,
Mindaugas
ode it goes
phone->kannel->proxy(the same)->MMSC.
Maybe kannel is passing only one chunk? Because I'm sending
~8kb MMS and proxy receives 'Content-length: 2306' HTTP POST.
So obviously something is lost.
What to do or how to track problem more precisely?
Thanks,
Mindaugas
oblems with Apache2 and chunked encoding POST. As I
wrote in original message if I used HTTP mode of posting MMS everything
worked fine even if transfer encoding was chunked. Problems appeared in
WSP mode when MMSes have to pass kannel.
Mindaugas
nly one chunk? Because I'm sending
~8kb MMS and proxy receives 'Content-length: 2306' HTTP POST.
So obviously something is lost.
What to do or how to track problem more precisely?
Thanks,
Mindaugas
data. :) And I do not know
yet how to get information about kernel samples.
Mindaugas
# op_time | tail -10
22006 0.1604 0. /lib/jbd.o
34394 0.2507 0. /usr/lib/libxml2.so.2.5.10
40080 0.2922 0.
/lib/modules/2.4.21-15.0.3.TLsmp/kernel/drivers/net/tg3.o
48515
soon afterwards.
How to patch Kannel to avoid coredump in such case? Clearly
we have to add:
if ( node == NULL ) return(-1);
And also some check for ret value in line 376 where we are calling
parse_document().
Something else?
Mindaugas
egfaulting.
It has other problems though. Not properly handling chunked encoding
posts and stopping from time to time.
Mindaugas
ain () from /lib/i686/libc.so.6
Mindaugas
uot; are
just informational or means something not good?
Mindaugas
==25725== ERROR SUMMARY: 0 errors from 0 contexts (suppressed: 14 from 3)
==25725== malloc/free: in use at exit: 552076 bytes in 6957 blocks.
==25725== malloc/free: 86189 allocs, 79232 frees, 9400491 bytes allocated.
==25725== For count
Hello,
I'm looking into backtrace of core dump and cannot
understand - how queue (wap/wtp_init.c) became NULL
(call to list_produce)?
Mindaugas
#0 0x080b9900 in lock (list=0x0) at gwlib/list.c:487
No locals.
#1 0x080b8f9e in list_append (list=0x0, item=0x45e309c8) at
gwlib/list.
, creating an error pdu
[0] WARNING: WTP PDU with unknown type -1
Mindaugas
P.S. I'm "from Internet" so I do not have good understaning about
how eveything works.
> Can you try cvs (if you do not use it already). I think
> this bug is fixed.
>
> Btw, PPG is part of wapbox, it implements *wap* push.
I'm using the latest cvs version. wap push is not used.
wap gateway is used by users just to browse.
Mindaugas
> Can you send your configuration file ?
Both wapbox and bearer box are started:
*box -v 4 wap.conf
Mindaugas
group = core
admin-port = 13000
wapbox-port = 13002
admin-password = x
status-password = x
admin-deny-ip = "*.*.*.*"
admin-allow-ip = "127.0.
> You should not have ppg running at all, then.
> Not having ppg configured should disable it.
> Can you tell your configuration flags ?
CFLAGS="-g -ggdb3" ./configure --with-malloc=native \
--disable-assertions --enable-cookies --enable-localtime --disable-sms
Mindaugas
Hello,
Follow up:
After I tried to run with following patch wapbox crashed with different
backtrace.
Both patch and gdb backtrace at the end of e-mail. No ideas why the crash
happened?
Mindaugas
#0 0x08080117 in wap_event_assert (event=0x0) at wap_events.c:222
No locals.
#1
just Kannel user who wants to resolve his problem and
is ready to assist in every possible way :).
Thanks,
Mindaugas
following code:
event = unpack_wdp_datagram_real(subdgram);
wap_event_assert(event);
And unpack_wdp_datagram_read returns NULL if it detects
truncated datagram. Without assertions the wapbox crashes
by trying to access event->type in wap/wap_events.c line 222.
How to fix that?
Mindaugas
. Second - what function
can I use to dump that datagram?
Thanks,
Mindaugas
rgs), char *);
if (format->has_prec && format->prec < (long) strlen(s))
n = format->prec;
else
==2139==>n = (long) strlen(s);
Mindaugas
(gdb) bt full
#0 0x002a95edac10 in strlen () from /lib64/tls/libc.so.6
No symbol table info available.
24 matches
Mail list logo
