Re: amavisd-new start/stop
Robert Millan wrote:
>>I wouldn't be surprised if other perl-based daemons have this same problem.
>
>
> AFAIK the kernel is in charge to interpret shebang headers and run the
> corresponding interpreter, so it wouldn't be strange that Linux and kFreeBSD
> differ in some way.
>
> How does this script's shebang look like?
I almost forgot about my own thread :)
robin@eddy:~$ head -1 /usr/sbin/amavisd-new
#!/usr/bin/perl -T
On a (just yesterday installed) FreeBSD 5.4 server I get (the binary is
/usr/sbin/amavisd):
vscan 45582 18.0 3.5 37908 36236 ?? S 2:08PM 0:03.35 amavisd
(ch8-avail) (perl5.8.6)
vscan 45595 5.8 3.5 37728 36496 ?? S 2:08PM 0:00.96 amavisd
(ch5-45595-05-3-idle) (perl5.8.6)
vscan 45590 5.6 3.4 37204 36000 ?? S 2:08PM 0:01.38 amavisd
(ch6-45590-06-2-idle) (perl5.8.6)
vscan 45580 3.8 3.5 38124 36772 ?? S 2:08PM 0:01.29 amavisd
(ch9-avail) (perl5.8.6)
vscan 501 0.0 3.3 35608 34488 ?? Ss 3:52PM 0:23.25 amavisd
(master) (perl5.8.6)
On my Debian GNU/kFreeBSD box I get (the binary is /usr/sbin/amavisd-new):
amavis 69243 0.0 0.0 26812 0 ? S Mar09 0:00
/usr/bin/perl -T /usr/sbin/amavisd-new start
amavis 69124 0.0 0.0 26968 0 ? S Mar09 0:00
/usr/bin/perl -T /usr/sbin/amavisd-new start
amavis 432 0.0 0.0 25908 0 ? Ss Mar09 0:00
/usr/bin/perl -T /usr/sbin/amavisd-new start
On a up-to-date Ubuntu box I get (the binary is /usr/sbin/amavisd-new):
amavis 12989 0.0 1.8 17864 14732 ? Ss 14:07 0:00 amavisd
(master)
amavis 12990 0.0 1.8 17864 14736 ? S 14:07 0:00 amavisd
(virgin child)
amavis 12991 0.0 1.8 17864 14732 ? S 14:07 0:00 amavisd
(virgin child)
The amavisd-new binary on Ubuntu seems to tell the system that it's
process name is 'amavisd <something>'. I assume the kFreeBSD package is
mostly the same as Ubuntu's, so amavisd-new on kFreeBSD somehow fails to
tell the kernel it's process name, or the kernel does not listen.
Is my logic going in the right direction?
I remember some time ago having done some programming in C and calling a
function to set the process name. But this stuff is in perl so there's
probably a perl function for it, which fails.
Still going in the right direction?
Robin
Reply to: