when i remove the
t_reply("404","not found")
in failure_route, everything works alright.
note: i intend to use the t_reply if there are no more branches available,
instead of a self generated 408
then, putting a t_release() on event_route[dialog:failed] got rid of the
error
WARNING: tm [t_
Hello,
do you get also a log message like:
CRITICAL: dialog [dlg_hash.c:794]: log_next_state_dlg(): bogus event 4
in state 5 for dlg ... ?
The warning itself should be harmless. Maybe you can grab all the log
messages with debug=3, that will help to get more hits about what happens.
Cheers
Hello,
I kind of forgot about this thread ...
The backtrace is clean and it doesn't show like a block/infinite loop
when you took it. Are you sure you took it when kamailio used 100%CPU?
How many times you run 'n' after backtrace? The loop in cbp could have
tens of iterations.
Maybe you ca
Hi guys,
was this issue resolved?
I encountered this issue also (4.0.5), in failure_route, i used
t_reply("404","not found")
never sent out, instead i get this in logs,
WARNING: tm [t_lookup.c:1564]: t_unref(): WARNING: script writer didn't
release transaction
Kelvin Chua
On Tue, Nov 26, 2013 a
Hello,
the backtrace is ok for the moment. I will look over it and come back
with results.
Cheers,
Daniel
On 11/25/13 12:10 PM, Efelin Novak wrote:
Hi Daniel,
sorry it took me more than I expected. Is this sufficient? Meanwhile I
found out that this happens when fr_inv_timer triggers and d
Hi Daniel,
sorry it took me more than I expected. Is this sufficient? Meanwhile I
found out that this happens when fr_inv_timer triggers and dialog module is
turned on.
Backtrace:
bt
#0 futex_release (lock=0xb4b90798) at ../../mem/../futexlock.h:137
#1 next_state_dlg (dlg=dlg@entry=0xb4bc5a38,
Hello,
I will investigate more -- meanwhile had some traveling. It would speed
up if you can send the backtrace of one process that blocks when you
applied the patch.
You need to connect with gdb to it:
gdb /path/to/kamailio _PID_
replace _PID_ with the pid of blocked kamailio process.
Che
Hi Daniel,
thanks for a reply. I applied the patch and now the Kamailio just prints
WARNING: tm [t_lookup.c:1564]: t_unref(): WARNING: script writer didn't
release transaction
and than freezes without any log. It does not resend the incoming "winning"
failure reply neither response to any other
Hello,
can you try attached patch?
Let me know if all goes fine and I will commit it to the repository.
Cheers,
Daniel
On 11/15/13 10:25 AM, Efelin Novak wrote:
Hi,
when I use t_reply("505", "Error"); in my failure route, the response
is not forwarded and following is written into a log:
Hi,
when I use t_reply("505", "Error"); in my failure route, the response is
not forwarded and following is written into a log:
kamailio[26216]: WARNING: tm [t_lookup.c:1559]: t_unref(): WARNING: script
writer didn't release transaction
plus next line is written exactly 416000 times into a log a
10 matches
Mail list logo
