Fail2ban and login_trusted_networks

[Javi Legido]

Hi there.

Tom, thanks for your reply. Indeed webmail (roundcube) has its own
mechanism for consecutive wrong logins, but I would like to have everything
centralized in one place (fail2ban in the container running dovecot).

Aki, thanks for your reply. Everything is working as expected, webmail
(roundcube) is sending the public IP address (8.8.8.8 in this example) to
dovecot, which is logging it like:

```
Dec 21 13:33:50 mail dovecot: imap-login: Login: user=<user1 at example.com>,
method=PLAIN, rip=8.8.8.8, lip=10.163.18.3, mpid=17151, TLS,
session=<oXIwgvm2fo9fF1Jz>
```



On Mon, 21 Dec 2020 at 13:20, Aki Tuomi <aki.tuomi at open-xchange.com> wrote:

> Dovecot can log client IP instead of connection IP, when webmail passes
> this information over using IMAP `ID` command, and webmail server has been
> added to login_trusted_networks.
>
> The keywords used for this feature are
>
> x-originating-ip
> x-originating-port
> x-connected-ip
> x-connected-port
> x-proxy-ttl (hop count)
>
> Aki
>
> > On 21/12/2020 14:12 Tom Hendrikx <tom at whyscream.net> wrote:
> >
> >
> > Hi,
> >
> > Ideally the webmail has it's own logfile, where it also emits error
> > messages containing the ip-address of the failed login attempt. This
> > could be as simple as a HTTP 401 error in the nginx/apache logfile on
> > the webmail domain. You can then instruct fail2ban to read that logfile
> > and disallow access to the webmail for the ip address.
> >
> > In the end, the attempts try to access the webmail, and not the IMAP
> > server directly. So it's better to block access to the webmail/web
> server.
> >
> > Kind regards,
> >       Tom
> >
> > On 21-12-2020 11:16, Javi Legido wrote:
> > > Hi there.
> > >
> > > First of all many thanks to all the people involved in this project
> for
> > > their time, I really appreciate it.
> > >
> > > Second my use case:
> > >
> > > a) Container running Webmail (roundcube) with dovecot-ident plugin
> > > enabled
> > > <
> https://github.com/roundcube/roundcubemail/issues/5336#issuecomment-228131074
> >.
> > > b) Container running Dovecot  2.3.4.1 (docker-mailserver-mysql
> > > <https://github.com/Kedu-SCCL/docker-mailserver-mysql>) with fail2ban
> > > enabled
> > >
> > > Since I need to add the private IP address of the webmail to
> > > "login_trusted_networks" to "...allow to override their IP addresses
> and
> > > ports" I can keep login to webmail even though if the IP is blocked.
> > >
> > > Question: there's any way to:
> > >
> > > a) Allow a certain IP range to override it's IP address and ports (as
> in
> > > "login_trusted_networks") but
> > > b) Be blocked, as any other incoming connections, by fail2ban?
> > >
> > > More context. Once the public IP is banned (8.8.8.8 in this example):
> > >
> > > ```
> > > 2020-12-21 10:10:31,371 fail2ban.filter         [309]: INFO
>  [dovecot]
> > > Found 8.8.8.8 - 2020-12-21 10:10:31
> > > 2020-12-21 10:10:39,189 fail2ban.filter         [309]: INFO
>  [dovecot]
> > > Found 8.8.8.8 - 2020-12-21 10:10:39
> > > 2020-12-21 10:10:51,222 fail2ban.filter         [309]: INFO
>  [dovecot]
> > > Found 8.8.8.8 - 2020-12-21 10:10:51
> > > 2020-12-21 10:10:52,008 fail2ban.actions        [309]: NOTICE
>  [dovecot]
> > > Ban 8.8.8.8
> > > ```
> > >
> > > I can't reach dovecot by telnet from this public IP:
> > >
> > > ```
> > > telnet mail.example.com <http://mail.example.com> 143
> > > Trying 9.9.9...
> > > telnet: Unable to connect to remote host: Connection refused
> > > ```
> > >
> > > Unless I removed the ban:
> > >
> > > ```
> > > docker exec mail fail2ban-client set dovecot unbanip 8.8.8.8
> > > ```
> > >
> > > Many thanks.
> > >
> > > Javier
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://dovecot.org/pipermail/dovecot/attachments/20201221/50143f13/attachment-0001.html>