[sr-dev] deamonize.c and the option DSYSLOG_ASYNC Kamailio 1.5

[Torben Friese]

Hi Henning,

I have found out that this NOT a lenny problem. It is general.

Have you seen the other post :

####

If I set the debug(in cfg) to 4 the proxy write logging messages, then
via 
$kamctl debug 2 
is everything fine(still logging). 

If the debug is set to 2 and I run the proxy in foreground I also
receive log messages. 

#####


Another interesting thing is if you bring in a log at line 205
(daemonize.c) it is working:

        LM_ERR("what ever \n");
	for (r=3; r < 32; r++){
		close(r);
	}

It seems that close() is not able to close the fd for the log (return
value from mksock in syslog_async.c). 

But If I let the for-loop counter start at 4 it is still not working.

After that I tried to set up log_fd as global( int log_fd=-1) and
include the headers, to check the "r" against the log_fd before execute
close(). But the same result.

Also a correct close(closelog_async()) in deamonize do not work. 

I am open for any ideas, maybe I have oversight something. Could you
reproduce this ? 


best regards 

Torben 







Am Donnerstag, den 26.11.2009, 12:40 +0100 schrieb Henning Westerholt:
> On Freitag, 20. November 2009, Torben Friese wrote:
> > OK sorry, the variable "r" is re-set in the for-loop, my
> fault(better
> > think before write ;-) ).
> >
> > But the loop is the problem in my opinion.
> >
> > If I put a LM_* log direct in fron of the close(line208 in
> daemonize.c),
> > everything seems fine.
> > Which means the close command fails (maybe at the right fd).
> 
> 
> 
> Hi Torben,
> 
> 
> 
> did not dig in the code in question so far, but is this still the same
> problem you reported on the user list, that it work on etch but not on
> lenny?
> 
> 
> 
> Regards,
> 
> 
> 
> Henning
>