Tony Finch wrote:
> On Tue, 21 Dec 2004, Tommy Butler wrote:
>
>>Anybody ever have problems like this?
>>
>>When trying to send email, I get this error in /var/log/exim4/rejectlog:
>>
>>2004-12-21 08:23:44 plain_saslauthd authenticator failed for ([192.168.0.111])
>>[69.15.114.65]: 435 Unable to authenticate at present (set_id=tommy): cannot
>>connect to saslauthd daemon at /var/run/saslauthd/mux: Permission denied
>
> Check that the exim user has permission to read the directory
> /var/run/saslauthd/.
WEIRD. Get this:
First I had to make sure that Debian-exim was in the sasl group:
$ usermod -G sasl Debian-exim
$ groups Debian-exim
Debian-exim : Debian-exim sasl
Then I checked permissions on /var/run/saslauthd/ as was suggested.
root@noot:~# stat /var/run/saslauthd/
File: `/var/run/saslauthd/'
Size: 136 Blocks: 0 IO Block: 131072 directory
Device: 303h/771d Inode: 43645 Links: 2
Access: (0710/drwx--x---) Uid: ( 0/ root) Gid: ( 1002/ sasl)
Access: 2004-12-21 08:42:05.294919056 -0600
Modify: 2004-12-21 08:39:28.101816040 -0600
Change: 2004-12-21 08:39:28.101816040 -0600
hmmmmmm. It looks like g+r needs to happen there. Right? Well so I...
$ chmod g+r /var/run/saslauthd/
...And all is well. No more errors or SMTP rejects. Everything works as
expected now. But I can't help but wonder why group read access wasn't already
on that folder. Well, at any rate I decide to restart the sasl auth daemon for
good measure, and much to my surprise, something very WEIRD happens...
root@noot:~# /etc/init.d/saslauthd restart
Stopping SASL Authentication Daemon: saslauthd.
Starting SASL Authentication Daemon: mode of `/var/run/saslauthd' changed to
0710 (rwx--x---)
saslauthd.
WOW. Now the errors come back and nothing works again. (I get the same
permission denied messages in my rejectlog and can't send mail.)
So why is this happening? What should I do? Obviously setting up a cron job to
constantly check if group has read access on /var/run/saslauthd/ is a stupid
hack. I'm stumped. Any thoughts?
--
Tommy Butler
tommy@???
