Hello,

    Strange error... ENOENT means that the file does not exists... That's
strange .. Well at least when doing an open in 4.3+BSD (see Stevens
"Advanced programming in the Unix environment) if you get that is because
the file does not exists... Did you compiled Postfix yourself? Do you have
any log of the compilation? I don't know if that has anything to do with the
problems you are experiencing but just try to guess what's wrong.

    I ask because I am running it (the same version) on a Solaris 8 box
(sparc) without any problem (it has been running for 21 days).

    Cheers

/B
----- Original Message -----
From: "goudy" <Gabriel.Goudycr-aquitaine.fr>
To: "Postfix users" <postfix-userspostfix.org>
Sent: Monday, April 02, 2001 4:34 PM
Subject: Re: Watchdog timeout

> I'm obliged to use single user mode because the boot is stopping and i
> can't do anything when it try to run Postfix.
>
> I use truss -p and i see somme errors :
>
> poll(0x08046F64, 1, 3600000) Err#91 ERESTART
> ...
> open("/usr/share/lib/zoneinfo/WET", 0_RDONLY) Err#2 ENOENT
> ...
> open("/var/run/syslog_door", 0_RDONLY) Err#2 ENOENT
> ...
> ... fatal: watchdog timeout
> sigsuspend(0x08046C4C) Err#4 EINTR
>
> The first error seems to be a socket error but i don't know if it's very
> important.
> I don't understand the two others because i can read the files and with
> "ls -al" i can see :
> -rw-r--r-- 1 root bin ... /usr/share/lib/zoneinfo/WET
> Drw-r--r-- 1 root other ... /var/run/syslog_door
>
> The third one should be a consequence.
>
> does someone have some informations ?
>
> Gabriel GOUDY.
>
> Wietse Venema wrote:
> >
> > Postfix uses watchdog timers that kill a process when nothing
> > happens for an unreasonable amount of time. This is a safety measure
> > to prevent Postfix from becoming stuck on broken systems.
> >
> > It looks like you have a broken system. I routinely test Postfix
> > on Solaris, and I know that many sites run Postfix on Solaris.
> > It's therefore most likely to be a site-specific problem.
> >
> > To find out what operation blocks, instrument your pickup server
> > with the truss system call tracer as per instructions in the
> > DEBUG_README file.
> >
> > BTW you don't need single user mode just to view the syslog file.
> >
> > Wietse
> >
> > goudy:
> > > Hi,
> > >
> > > I have just installed Postfix on my server and now, i can't boot.
> > > The boot stop when it try to run postfix :
> > >
> > > postfix-script: starting the Postfix mail system
> > >
> > > and after a few minutes
> > >
> > > mail postfix/pickup[365]: fatal: watchdog timeout
> > >
> > > I just can boot in single mode (after savage reboot), and i can read
on
> > > the log something like that :
> > >
> > > mail postfix/master[15056]: warning: /usr/libexec/postfix/pickup: bad
> > > command startup -- throttling
> > > mail postfix/pickup [ ]: fatal: watchdog timeout
> > > ...
> > >
> > > (and nothing more about socket or other things)
> > >
> > > Others informations i have :
> > >
> > > Solaris 8
> > > Postfix release-20010228
> > > openLDAP 2.0.7 (But actually, I just install postfix and test
so
> > > I don't use LDAP with
> > > it.)
> > >
> > > Does anyone know a solution ?
> > >
> > > Thanks.
> > >
> > > PS : Another point, I choose to revoke maildrop directory world write
> > > permission and to enable set-gid privileges on a small "maildrop"
> > > command. It's may be important.
> > >
> > >
> > > --
> > > Gabriel GOUDY
> > >
> > > -
> > > To unsubscribe, send mail to majordomopostfix.org with content
> > > (not subject): unsubscribe postfix-users
> > >
> > >
> >
> > -
> > To unsubscribe, send mail to majordomopostfix.org with content
> > (not subject): unsubscribe postfix-users
>
> -
> To unsubscribe, send mail to majordomopostfix.org with content
> (not subject): unsubscribe postfix-users
>
>

-
To unsubscribe, send mail to majordomopostfix.org with content
(not subject): unsubscribe postfix-users

• Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]