[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: syslog-ng



Okay, I've got this narrowed down to a problem with syslog-ng which
doesn't die when asked to, it just enters said 'every other time' state.
It explains a lot of things, including why the PID doesn't change when
restarting, I was assuming it only SIGHUPed it, big mistake.

Now, I've attached strace to it when it's still alive and got:

freki:~# /etc/init.d/syslog-ng start
Starting system logging: syslog-ng.
freki:~# ps aux | grep syslog-ng | grep -v grep
root      1777  0.0  0.6  1600  784 ?        S    10:02   0:00 /sbin/syslog-ng
freki:~# strace -p 1777
time(NULL)                              = 1034406255
poll([{fd=6, events=0}, {fd=5, events=POLLIN}, {fd=3, events=POLLIN}], 3, 100) = 0
poll(

Now, in another terminal, kill 1777, the result in the straced terminal:
freki:~# strace -p 1777
time(NULL)                              = 1034406255
poll([{fd=6, events=0}, {fd=5, events=POLLIN}, {fd=3, events=POLLIN}], 3, 100) = 0
poll([{fd=6, events=0}, {fd=5, events=POLLIN}, {fd=3, events=POLLIN}], 3, 582000) = -1 EINTR (Interrupted system call)
--- SIGTERM (Terminated) ---
freki:~# ps aux | grep syslog-ng | grep -v grep
root      1777  0.0  0.6  1600  784 ?        S    10:02   0:00 /sbin/syslog-ng

If I do any of the things mentioned in the previous email, it dies, if
not, it just sits there.
Not being a programmer, I can't really make much of that, I can tell you
that it happens on another Sid box I've tested it on though. I'm also
guessing this is a bug and that I should file it, just waiting to see if
I haven't missed something silly, like at first (heh, the PID thing was
obvious, now that I look back a it).

On Sat, 2002-10-12 at 08:06, Birzan George Cristian wrote:
> Hello
> 
> I'm running Unstable and I am having some very (IMO) weird trouble with
> syslog-ng. In a nutshell, it just dies when I try to log anything to it,
> every other time.... Long story, including 'logs' follows:
> 
> freki:~# /etc/init.d/syslog-ng restart
> Stopping system logging: syslog-ng.
> Starting system logging: syslog-ng.
> freki:~# ps aux | grep syslog-ng | grep -v grep
> root      1468  0.0  0.6  1600  784 ?        S    07:50   0:00 /sbin/syslog-ng
> freki:~# logger willdie
> freki:~# ps aux | grep syslog-ng | grep -v grep
> freki:~# /etc/init.d/syslog-ng start
> Starting system logging: syslog-ng.
> freki:~# ps aux | grep syslog-ng | grep -v grep
> root      1482  0.0  0.6  1600  784 ?        S    07:50   0:00 /sbin/syslog-ng
> freki:~# logger willnotdie
> freki:~# ps aux | grep syslog-ng | grep -v grep
> root      1482  0.0  0.6  1608  828 ?        S    07:51   0:00 /sbin/syslog-ng
> 
> As if this weren't enough, it seems to blindly obey Murhpy's laws as it
> will not die if started in debugging mode. On the other hand, it will
> die as soon as I attach strace to it, again, every other time...
> 
> freki:~# ps aux | grep syslog-ng | grep -v grep
> root      1482  0.0  0.6  1608  828 ?        S    07:50   0:00 /sbin/syslog-ng
> freki:~# strace -p 1482
> time(NULL)                              = 1034398350
> poll([{fd=11, events=0}, {fd=10, events=0}, {fd=9, events=0}, {fd=8, events=0}, {fd=7, events=0}, {fd=6, events=0}, {fd=5, events=POLLIN}, {fd=3, events=POLLIN}], 8, 100) = 0
> poll(
> (waits for something to happen, I'd say, as is normal.)
> freki:~# /etc/init.d/syslog-ng restart
> Stopping system logging: syslog-ng.
> Starting system logging: syslog-ng.
> freki:~# ps aux | grep syslog-ng | grep -v grep
> root      1482  0.0  0.6  1608  828 ?        S    07:51   0:00 /sbin/syslog-ng
> freki:~# strace -p 1482
> _exit(0)                                = ?
> 
> This behaviour is consistent, tried about 30 times so far, only dies
> every other time...
> Using the Sid version of it, 1.5.20-1, the rest of the system is fairly
> up-to-date, dist-upgraded about 12 hours ago
> 
> My question is, being at a tolal loss, what else can I try?
-- 
PGP public key: http://www.wolfheart.ro/pubkey.asc

Attachment: signature.asc
Description: This is a digitally signed message part


Reply to: