summaryrefslogtreecommitdiffstats
path: root/roles/MX
Commit message (Collapse)AuthorAgeFiles
* Prefix ‘ipaddr’ and ‘ipv4’ with ‘ansible.utils.’.Guilhem Moulin2022-10-112
| | | | | | | This silences the following deprecation warning: Use 'ansible.utils.ipaddr' module instead. This feature will be removed from ansible.netcommon in a release after 2024-01-01. Deprecation warnings can be disabled by setting deprecation_warnings=False in ansible.cfg.
* wibbleGuilhem Moulin2020-05-181
|
* MX: Port to Debian 10.Guilhem Moulin2020-05-162
| | | | | | | | For postfix, don't defer if "abused legit". (I.e., DBL return code in the 127.0.1.100+ range.) This used to work for Postfix 3.1.14 (Stretch) but for 3.4.8 (Buster) the 'defer_if_reject' also applies to $smtpd_relay_restrictions, to reject_unauth_destination & reject_unlisted_recipient in particular.
* MX: Install OpenDMARC to add Authentication-Results headers.Guilhem Moulin2020-05-166
| | | | | | | | On the infrastructure boundary. We don't reject/quarantine as it would affect members who forward their mail sent to <user@example.com> to <user@fripost.org>. Members can install Sieve rules to send any messages with failed Authentication-Results headers directly in their spambox.
* MX: chroot postscreen(8), smtpd(8) and cleanup(8) daemons.Guilhem Moulin2018-12-097
| | | | | | Unlike what we wrote in 2014 (cf. 4fb4be4d279dd94cab33fc778cfa318b93d6926f) the postscreen(8) server can run chrooted, meaning we can also chroot the smtpd(8), tlsproxy(8), dnsblog(8) and cleanup(8) daemons.
* MX: don't override 5XY reject codes to 554.Guilhem Moulin2018-12-091
|
* postfix: remove explicit default 'mail_owner = postfix'.Guilhem Moulin2018-12-061
|
* Upgrade MX baseline to Debian Stretch.Guilhem Moulin2018-12-031
|
* Upgrade syntax to Ansible 2.7 (apt module).Guilhem Moulin2018-12-031
|
* Postfix: replace cdb & btree tables with lmdb ones.Guilhem Moulin2018-12-032
| | | | Cf. lmdb_table(5).
* Harden anti spam on the MX:es.Guilhem Moulin2018-06-094
|
* Postscreen: improve DNSBL sites and scores.Guilhem Moulin2018-04-041
|
* postscreen: lower zen.spamhaus.org DNSBL score from 3 to 2 on the MX:es.Guilhem Moulin2017-06-051
| | | | | So being listed on that BL doesn't yield a flat reject if the IP isn't also listed to other lists.
* Use blackhole subdomain for sender addresses of verify probes.Guilhem Moulin2017-05-161
| | | | | | | | | | | These addresses need to be accepted on the MX:es, as recipients sometimes phone back during the SMTP session to check whether the sender exists. Since a time-dependent suffix is added to the local part (cf. http://www.postfix.org/postconf.5.html#address_verify_sender_ttl) it's not enough to drop incoming mails to ‘double-bounce@fripost.org’, and it's impractical to do the same for /^double-bounce.*@fripost\.org$/.
* Change group of executables in /usr/local/{bin,sbin} from root to staff.Guilhem Moulin2017-05-141
|
* Allow SMTP client from whitelisted IPs to bypass postscreen checks.Guilhem Moulin2017-01-141
|
* Postscreen: Give temporary whitelist status to primary MX addresses only.Guilhem Moulin2016-09-201
|
* postfix: commit the master.cf symlinks.Guilhem Moulin2016-07-121
|
* Route all internal SMTP traffic through IPsec.Guilhem Moulin2016-07-102
|
* Postfix MX/MSA instances: put certs in the the instance's $config_directory.Guilhem Moulin2016-07-102
|
* Postfix MX/MSA instances: don't ask the remote SMTP client for a client ↵Guilhem Moulin2016-07-101
| | | | | | | certificate. See postconf(5). This avoids the “(Client did not present a certificate)” messages in the Received headers.
* Postfix: don't share the master.cf between the instances.Guilhem Moulin2016-07-102
|
* postfix: Don't explicitly set inet_interfaces=all as it's the default.Guilhem Moulin2016-07-101
|
* Change the pubkey extension from .pem to .pub.Guilhem Moulin2016-07-101
|
* certs/public: fetch each cert's pubkey (SPKI), not the cert itself.Guilhem Moulin2016-06-151
| | | | To avoid new commits upon cert renewal.
* postfix: Update to recommended TLS settings.Guilhem Moulin2016-05-181
| | | | | | | | Following Viktor Dukhovni's 2015-08-06 recommendation http://article.gmane.org/gmane.mail.postfix.user/251935 (We're using stronger ciphers and protocols in our own infrastructure.)
* postfix: unset 'smtpd_tls_session_cache_database'.Guilhem Moulin2016-05-181
| | | | | | Following Viktor Dukhovni's 2015-08-06 recommendation for Postfix >= 2.11 http://article.gmane.org/gmane.mail.postfix.user/251935
* Move /etc/ssl/private/dhparams.pem to /etc/ssl/dhparams.pem and make it public.Guilhem Moulin2016-05-181
| | | | | | | | | | Ideally we we should also increase the Diffie-Hellman group size from 2048-bit to 3072-bit, as per ENISA 2014 report. https://www.enisa.europa.eu/publications/algorithms-key-size-and-parameters-report-2014 But we postpone that for now until we are reasonably certain that older client won't be left out.
* postfix: disable weak ciphers for the 'encrypt' TLS security level.Guilhem Moulin2016-05-181
| | | | That is, on the MSA and in our local infrastructure.
* Add an ansible module 'fetch_cmd' to fetch the output of a remote command ↵Guilhem Moulin2016-05-181
| | | | | | locally. And use this to fetch all X.509 leaf certificates.
* Let's EncryptGuilhem Moulin2016-03-021
|
* Upgrade playbooks to Ansible 2.0.Guilhem Moulin2016-02-121
|
* Use the Let's Encrypt CA for our public certs.Guilhem Moulin2015-12-202
|
* Automatically fetch X.509 certificates, and add them to git.Guilhem Moulin2015-12-031
|
* Fix address verification probes on the MSA.Guilhem Moulin2015-09-161
| | | | | Put all relay restrictions under smtpd_relay_restrictions and leave smtpd_recipient_restrictions empty, since we don't do DNSBL.
* Don't bounce unverified recipients upon 4xx errors.Guilhem Moulin2015-06-111
| | | | | | | We don't want to bounce messages for which the recipient(s)' MTA replies 451 due to some greylisting in place. We would like to accept 451 alone, but unfortunately it's not possible to bounce unverified recipients due to DNS or networking errors.
* Configure munin nodes & master.Guilhem Moulin2015-06-102
| | | | | Interhost communications are protected by stunnel4. The graphs are only visible on the master itself, and content is generated by Fast CGI.
* Add a reserved domain 'discard.fripost.org' to discard messages.Guilhem Moulin2015-06-072
| | | | | ‘noreply@’ aliases can be added by routing them to ‘@discard.fripost.org’.
* Upgrade the MX configuration from Wheezy to Jessie.Guilhem Moulin2015-06-079
| | | | | | In particular, since Postfix is now able to perform LDAP lookups using SASL, previous hacks with simble binds on cn=postfix,ou=services,… can now be removed.
* logjam mitigation.Guilhem Moulin2015-06-071
|
* Configure the list manager (Sympa).Guilhem Moulin2015-06-071
|
* typoGuilhem Moulin2015-06-071
|
* typoGuilhem Moulin2015-06-071
|
* Split templates / files in lookup tables.Guilhem Moulin2015-06-078
|
* Replace Postgrey with postscreen.Guilhem Moulin2015-06-079
| | | | | | | | | | | See http://www.postfix.org/POSTSCREEN_README.html and http://rob0.nodns4.us/postscreen.html It's infortunate that smtpd(8) cannot be chrooted any longer, which means that we have to un-chroot cleanup(8) as well. Indeed, currently smtpd(8) uses $virtual_alias_maps for recipient validation; later cleanup(8) uses it again for rewriting. So these processes need to be both chrooted, or both not.
* Verify the validity of users before that of aliases.Guilhem Moulin2015-06-071
|
* Remove reject_unknown_sender_domain from the MX.Guilhem Moulin2015-06-071
| | | | | There are false-positive with that, for instead due to SOA records pointing to non-existing subdomains.
* wibbleGuilhem Moulin2015-06-071
|
* Use the raw 'fripostListManager' as routing internal subdomain.Guilhem Moulin2015-06-071
|
* Fix $smtpd_sender_restrictions.Guilhem Moulin2015-06-071
| | | | | | | | | | | | On the MDA the domain is our 'mda.fripost.org', there is no need to perform an extra DNS lookup. The MSA does not perform local or virtual delivery, but relays everything to the outgoing SMTP proxy. On the MX, there is no need to check for recipient validity as we are the final destination; but unsure that the RCPT TO address is a valid recipient before doing the greylisting.