One thing I should possibly also note: I'm using MailScanner, hence the 
sendmail service is started via ms-sendmail rather than sendmail.  That is, I'm 
using /usr/lib/systemd/system/ms-sendmail.service (and the associated 
ms-sendmail-in and ms-sendmail-out, which get loaded through this file), rather 
than the default sendmail.service...

I've changed the journalmatch to ms-sendmail.service but that doesn't seem to 
help at all.

Thanks in advance for assistance.

--- Amir

> On Mar 25, 2019, at 8:48 PM, Amir Caspi <[email protected]> wrote:
> 
> Hi all,
> 
>       I'm setting up a new CentOS 7 server with f2b 0.9.7 (from EPEL) and I'm 
> having trouble with one of my sendmail rules.  I set up a custom rule to ban 
> servers that fail SMTP AUTH and then hang up... while I would normally want 
> to ban these guys using the PAM failure, unfortunately the current cyrus-sasl 
> implementation means that saslauthd doesn't log the remote host, or send the 
> rhost info to PAM to log... so this is (right now) the only way to do it.
> 
> Unfortunately, although my filter matches just fine using fail2ban-regex, the 
> server is never triggering that filter.
> 
> The filter is sendmail-noauth.conf, as follows:
> failregex = ^%(__prefix_line)s\w{14}: (\S+ )?\[<HOST>\]( \(may be forged\))? 
> did not issue MAIL/EXPN/VRFY/ETRN during connection to MTA$
> 
> It's enabled in jail.local as follows, with default maxretry = 3, findtime = 
> 600, bantime = 600:
> [sendmail-noauth]
> enabled  = true
> port     = submission,465,smtp
> logpath  = %(syslog_mail)s
> backend  = %(syslog_backend)s
> 
> This is on CentOS 7, so I'm loading paths_fedora.conf which has syslog_mail = 
> /var/log/maillog and syslog_backend = systemd.  Default action is mwl so I 
> get the email notification.
> 
> /var/log/maillog has hundreds of lines from the same offender:
> Mar 26 02:40:53 servername sm-mta[10953]: x2Q2eiHC010953: 
> mta9.imxonlines.co.za [91.212.150.89] (may be forged) did not issue 
> MAIL/EXPN/VRFY/ETRN during connection to MTA
> 
> If I MANUALLY ban the IP using fail2ban-client, the resulting email is 
> looking up the correct lines from the correct logfile.
> 
> So, in short: fail2ban-regex matches just fine, and manual banning shows that 
> the correct logfile is being read.  But automated banning is not working.
> 
> This same rule works just fine on my old CentOS 5 box using f2b 0.8.14.
> 
> Can anyone help?  I have no idea why it's not banning when fail2ban-regex 
> works fine, as does manual banning.
> 
> Thanks!
> 
> --- Amir
> 

_______________________________________________
Fail2ban-users mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/fail2ban-users

Reply via email to