In Slackware Current you might notice output in /var/log/maillog like this
following the execution of scheduled cron (dcron) tasks. Fortunately it's very easy to gratify postdrop. As root, just execute
The error message goes away.

This does not appear to be an issue with stock installs of 14.2 Stable, only Current.

Postfix 2.10.0 changed /var/spool/postfix/public/pickup from a fifo to a socket to avoid needless disk spinups when the mail system is idle. Here on -current I see the appropriate socket when postfix is started:

Of course, you can still continue to use a fifo if you wish, but that's considered obsolete. You might want to take a look at your master.cf if you've made any changes to it or got it from someplace else. The pickup line ships containing "unix" which causes postfix to create the socket, but if it says "fifo" instead it might cause the behavior reported here.

3 members found this post helpful.

Thank you for this info.

My /etc/postfix/master.cf is "stock". I pulled a current ISO on alienbob's site in late January and did a fresh install, so I'm starting with postfix-3.4.8 and retaining none of my postfix config from previous installs. Also, the relevant line in /etc/postfix/master.cf is there:
and yet the socket file is not created (and hence the error message).

Now I'm unsure what's going on ...

kestralis,
I had exact same problem till I made /etc/rc.d/rc.postfix executable and then started it. Now I have same as Pat.
HTH

1 members found this post helpful.

So I needed only to start postfix one time to populate the folder with the needed sockets. Thanks for this.

I wonder if this step could be added to the installation or first-boot related scripts so no one else runs into this issue again? I'm sure we're not the only ones who don't set postfix to start at boot when doing a clean install.

I don't think the sockets are of any use if postfix isn't running.

1 members found this post helpful.

It might be logrotate related, on what time did this occour in the log?
What other logentries are before/after that?

Sorry to necropost this but just found this on my lap that is an ancient but fresh install of 15.0

Got 5 maillog files fulled with

This occurs 3 times every 20 minutes, 9 lines per hour.
Just created the fifo before I found this thread.

I bet that could get some more attention...

Edit :
Now I checked on a fresh VM the maillog file... empty. Will try to set a cronjob an see if I can reproduce the error.
Perhaps could try this
I bethat could get some more attention...

Unix systems are supposed to have a mail server. During many years, the mail server of choice was sendmail, nowadays postfix is the default mail server in Slackware.

The cron unix feature sends output from cron jobs by mail to the owner of the cron job. If you don't want such mails, you should make sure that all your cron jobs redirect their output (stdout and stderr) to /dev/null. If you want those mails, you will of course need a mail server running on the local machine. With no mail server running and cron jobs trying to send error messages or other output by mail you should instead expect to find error messages about lost mails in your log files.

If you really do not want any output from your cron jobs and no logs about failures to send such output you could make sure that crond is started with "-M /bin/true" which will simply use /bin/true to send mails and /bin/true allways reports success.

regards Henrik

2 members found this post helpful.

Thanks Henrik for your instructive reply.

Remaining drawback : the message is not informative (file do not exist instead of no mail server) and messages are lost if I redirect to /dev/null whereas I got them when I start postfix if not.

I understand I was misusing cron but still think this could be fixed.

Regards