Hi, First thank you for trying to help, but it seems that you missed a part: if I execute `postfix stop; postfix start` manually in the terminal, the problem is "solved" until the next reboot.
I am running in a real machine, and using real HDD not NFS. Interestingly when I try to flush the postqueue when the problem exists, a similar error is thrown: ~~~~~~ Jun 3 13:13:06 localhost postfix/flush[2705]: fatal: open lock file pid/unix.flush: cannot open file: Permission denied Jun 3 13:13:07 localhost postfix/master[1793]: warning: process /usr/libexec/postfix/flush pid 2705 exit status 1 Jun 3 13:13:07 localhost postfix/master[1793]: warning: /usr/libexec/postfix/flush: bad command startup -- throttling ~~~~~~ Which can also be solved by the same method. I don't believe this is really a permission problem, because I see that `/usr/libexec/postfix/master` is run as root even when the problem exists. Perhaps I will write once more: When the service is automatically started with the system, it fails to open lock files. If I manually start it in the terminal, everything is all right. Here's my `postconf -n`: ~~~~~~ alias_database = hash:/etc/aliases alias_maps = hash:/etc/aliases command_directory = /usr/sbin config_directory = /etc/postfix daemon_directory = /usr/libexec/postfix data_directory = /var/lib/postfix debug_peer_level = 2 html_directory = no inet_protocols = all mail_owner = postfix mailq_path = /usr/bin/mailq.postfix manpage_directory = /usr/share/man mydestination = $myhostname, localhost.$mydomain, localhost newaliases_path = /usr/bin/newaliases.postfix queue_directory = /var/spool/postfix readme_directory = /usr/share/doc/postfix-2.6.6/README_FILES sample_directory = /usr/share/doc/postfix-2.6.6/samples sendmail_path = /usr/sbin/sendmail.postfix setgid_group = postdrop unknown_local_recipient_reject_code = 550 virtual_alias_maps = hash:/etc/postfix/virtual virtual_gid_maps = static:5000 virtual_mailbox_base = /var/mail/vmailbox virtual_mailbox_domains = example.com foo.bar.com virtual_mailbox_maps = hash:/etc/postfix/vmailbox virtual_minimum_uid = 100 virtual_uid_maps = static:5000 ~~~~~~ Thanks, Alvin Wong 2012/6/4 Jeroen Geilman <jer...@adaptr.nl>: > On 06/03/2012 06:37 AM, Alvin Wong wrote: >> >> Hi, >> >> I am running postfix 2.6.6 on CentOS 6.2. >> I set up postfix to use virtual mailbox referring to this >> (http://www.postfix.org/VIRTUAL_README.html#virtual_mailbox) >> >> When the server starts, it can't deliver mail to virtual mailbox >> (mails are stuck in the queue). >> I see this in the maillog: >> >> ~~~~~~ >> Jun 2 20:35:43 localhost postfix/virtual[2122]: fatal: open lock file >> pid/unix.virtual: cannot open file: Permission denied >> Jun 2 20:35:44 localhost postfix/master[1784]: warning: process >> /usr/libexec/postfix/virtual pid 2122 exit status 1 >> Jun 2 20:35:44 localhost postfix/master[1784]: warning: >> /usr/libexec/postfix/virtual: bad command startup -- throttling >> ~~~~~~ >> >> However, when I execute `postfix stop; postfix start` in shell, >> postfix starts delivering mail to virtual mailbox and the message >> didn't appear again. >> But then when the machine is restarted (both warm reboot and cold >> reboot), the problem appears again. The only fix I know is to execute >> `postfix stop; postfix start` manually. >> >> The file: `/var/spool/postfix/pid/unix.virtual` is owned by root and >> has permission 0600 (rw for root only). It was set with `postfix >> set-permissions`. >> >> This is quite strange because when the server boots and start postfix >> automatically, it is supposed to do the same as `postfix start`. >> Anyone have any idea? > > > Is this inside a VM, or are you running the spool off NFS ? > Also check apparmor/SELinux for oddities. > If all else fails , examine the system for cron jobs that do weird stuff. > > You could also set up an inotify watch on the lock file to record who > changed it and when, but I am not too familiar with that - you will need to > read about it. > >> Thanks, >> Alvin Wong > > > > -- > J. >
