mta-sts-daemon.yml - Man Page

configuration file for mta-sts-daemon


This configuration file configures the listening socket, caching behaviour, and manipulation of MTA-STS mode.


The file is in YAML syntax with the following elements:

host: (str) daemon bind address. Default:

port: (int) daemon bind port. Default: 8461

path: (str) daemon UNIX socket bind address (path). If specified, host and port are ignored and UNIX socket is bound instead of TCP.

mode: (int) file mode for daemon UNIX socket. If not specified default filemode is used. This option has effect only when UNIX socket is used. If file mode specified in octal form (most common case), it has to be prepended with leading zero. Example: 0666

reuse_port: (bool) allow multiple instances to share same port (available on Unix, Windows). Default: true

cache_grace: (float) age of cache entries in seconds which do not require policy refresh and update. Default: 60

shutdown_timeout: (float) time limit granted to existing client sessions for finishing when server stops. Default: 20

  • type: (str: internal|sqlite|redis|redis_sentinel) cache backend type. Default: internal
  • options:
    • Options for internal type:

      • cache_size: (int) number of cache entries to store in memory. Default: 10000
    • Options for sqlite type:

      • filename: (str) path to database file
      • threads: (int) number of threads in pool for SQLite connections. Default: number of CPUs
      • timeout: (float) timeout in seconds for acquiring connection from pool or DB lock. Default: 5
    • Options for redis type:

      • All parameters are passed to aioredis.from_url [0]. Check there for a parameter reference.
    • Options for redis_sentinel type:

      • sentinel_master_name: (str) name of the sentinel master
      • sentinels: (list)(tuple) list of sentinels in form of ip/fqdn and port
      • All other parameters are passed to aioredis.sentinel.Sentinel [1]. For additional details check [2].
  • enabled: (bool) enable proactive policy fetching in the background. Default: false
  • interval: (int) if proactive policy fetching is enabled, it is scheduled every this many seconds. It is unaffected by cache_grace and vice versa. Default: 86400
  • concurrency_limit: (int) the maximum number of concurrent domain updates. Default: 100
  • grace_ratio: (float) proactive fetch for a particular domain is skipped if its cached policy age is less than interval/grace_ratio. Default: 2.0
  • strict_testing: (bool) enforce policy for testing domains. Default: false
  • timeout: (int) network operations timeout for resolver in that zone. Default: 4
  • require_sni: (bool) add option servername=hostname to policy responses to make Postfix send SNI in TLS handshake as required by RFC 8461. Requires Postfix version 3.4+. Default: true
    • Same as options in default_zone

The timeout is used for the DNS and HTTP requests.

MTA-STS "testing" mode can be interpreted as "strict" mode.  This may be useful (though noncompliant) in the beginning of MTA-STS deployment, when many domains operate under "testing" mode.


port: 8461
reuse_port: true
shutdown_timeout: 20
  enabled: true
  interval: 86400
  concurrency_limit: 100
  grace_ratio: 2
  type: internal
    cache_size: 10000
  strict_testing: false
  timeout: 4
    strict_testing: false
    timeout: 4

See Also

mta-sts-daemon(1), mta-sts-query(1)





Referenced By

mta-sts-daemon(1), mta-sts-query(1).

2024-01-25 postfix-mta-sts-resolver