JAIL.CONF(5) Fail2Ban Configuration JAIL.CONF(5)

jail.conf - configuration for the fail2ban server

fail2ban.conf fail2ban.d/*.conf fail2ban.local fail2ban.d/*.local

jail.conf jail.d/*.conf jail.local jail.d/*.local

action.d/*.conf action.d/*.local action.d/*.py

filter.d/*.conf filter.d/*.local

Fail2ban has four configuration file types:

Fail2Ban global configuration (such as logging)
Filters specifying how to detect authentication failures
Actions defining the commands for banning and unbanning of IP address
Jails defining combinations of Filters with Actions.

*.conf files are distributed by Fail2Ban. It is recommended that *.conf files should remain unchanged to ease upgrades. If needed, customizations should be provided in *.local files. For example, if you would like to enable the [ssh-iptables-ipset] jail specified in jail.conf, create jail.local containing

[ssh-iptables-ipset]

enabled = true

In .local files specify only the settings you would like to change and the rest of the configuration will then come from the corresponding .conf file which is parsed first.

In addition to .local, for jail.conf or fail2ban.conf file there can be a corresponding .d/ directory containing additional .conf files. The order e.g. for jail configuration would be:

jail.conf
jail.d/*.conf (in alphabetical order)
jail.local
jail.d/*.local (in alphabetical order).

i.e. all .local files are parsed after .conf files in the original configuration file and files under .d directory. Settings in the file parsed later take precedence over identical entries in previously parsed files. Files are ordered alphabetically, e.g.

fail2ban.d/01_custom_log.conf - to use a different log path

jail.d/01_enable.conf - to enable a specific jail
jail.d/02_custom_port.conf - to change the port(s) of a jail.

Configuration files have sections, those specified with [section name], and name = value pairs. For those name items that can accept multiple values, specify the values separated by spaces, or in separate lines space indented at the beginning of the line before the second value.

Configuration files can include other (defining common variables) configuration files, which is often used in Filters and Actions. Such inclusions are defined in a section called [INCLUDES]:

indicates that the specified file is to be parsed before the current file.
indicates that the specified file is to be parsed after the current file.

Using Python "string interpolation" mechanisms, other definitions are allowed and can later be used within other definitions as %(name)s.

Fail2ban has more advanced syntax (similar python extended interpolation). This extended interpolation is using %(section/parameter)s to denote a value from a foreign section.
Besides cross section interpolation the value of parameter in [DEFAULT] section can be retrieved with %(default/parameter)s.
Fail2ban supports also another feature named %(known/parameter)s (means last known option with name parameter). This interpolation makes possible to extend a stock filter or jail regexp in .local file (opposite to simply set failregex/ignoreregex that overwrites it), e.g.

baduseragents = IE|wget|%(my-settings/baduseragents)s
failregex = %(known/failregex)s
            useragent=%(baduseragents)s

Additionally to interpolation %(known/parameter)s, that does not works for filter/action init parameters, an interpolation tag <known/parameter> can be used (means last known init definition of filters or actions with name parameter). This interpolation makes possible to extend a parameters of stock filter or action directly in jail inside jail.conf/jail.local file without creating a separately filter.d/*.local file, e.g.

# filter.d/test.conf: 
[Init]
test.method = GET
baduseragents = IE|wget
[Definition]
failregex = ^%(__prefix_line)\s+"<test.method>"\s+test\s+regexp\s+-\s+useragent=(?:<baduseragents>)
# jail.local:
[test]
# use filter "test", overwrite method to "POST" and extend known bad agents with "badagent":
filter = test[test.method=POST, baduseragents="badagent|<known/baduseragents>"]

Comments: use '#' for comment lines and '; ' (space is important) for inline comments.

The items that can be set in section [Definition] are:

verbosity level of log output: CRITICAL, ERROR, WARNING, NOTICE, INFO, DEBUG, TRACEDEBUG, HEAVYDEBUG or corresponding numeric value (50-5). Default: INFO (equal 20)
log target: filename, SYSLOG, STDERR or STDOUT. Default: STDOUT if not set in fail2ban.conf/fail2ban.local
Note. If fail2ban running as systemd-service, for logging to the systemd-journal, the logtarget could be set to STDOUT
Only a single log target can be specified. If you change logtarget from the default value and you are using logrotate -- also adjust or disable rotation in the corresponding configuration file (e.g. /etc/logrotate.d/fail2ban on Debian systems).
socket filename. Default: /var/run/fail2ban/fail2ban.sock
This is used for communication with the fail2ban server daemon. Do not remove this file when Fail2ban is running. It will not be possible to communicate with the server afterwards.
PID filename. Default: /var/run/fail2ban/fail2ban.pid
This is used to store the process ID of the fail2ban server.
option to allow IPv6 interface - auto, yes (on, true, 1) or no (off, false, 0). Default: auto
This value can be used to declare fail2ban whether IPv6 is allowed or not.
Database filename. Default: /var/lib/fail2ban/fail2ban.sqlite3
This defines where the persistent data for fail2ban is stored. This persistent data allows bans to be reinstated and continue reading log files from the last read position when fail2ban is restarted. A value of None disables this feature.
Max number of matches stored in database per ticket. Default: 10
This option sets the max number of matched log-lines could be stored per ticket in the database. This also affects values resolvable via tags <ipmatches> and <ipjailmatches> in actions.
Database purge age in seconds. Default: 86400 (24hours)
This sets the age at which bans should be purged from the database.

The config parameters of section [Thread] are:
Stack size of each thread in fail2ban. Default: 0 (platform or configured default)
This specifies the stack size (in KiB) to be used for subsequently created threads, and must be 0 or a positive integer value of at least 32.

The following options are applicable to any jail. They appear in a section specifying the jail name or in the [DEFAULT] section which defines default values to be used if not specified in the individual section.

name of the filter -- filename of the filter in /etc/fail2ban/filter.d/ without the .conf/.local extension.
Only one filter can be specified.
filename(s) of the log files to be monitored, separated by new lines.
Globs -- paths containing * and ? or [0-9] -- can be used however only the files that exist at start up matching this glob pattern will be considered.

Optional space separated option 'tail' can be added to the end of the path to cause the log file to be read from the end, else default 'head' option reads file from the beginning

Ensure syslog or the program that generates the log file isn't configured to compress repeated log messages to "*last message repeated 5 time*s" otherwise it will fail to detect. This is called RepeatedMsgReduction in rsyslog and should be Off.

encoding of log files used for decoding. Default value of "auto" uses current system locale.
Force the time zone for log lines that don't have one.

If this option is not specified, log lines from which no explicit time zone has been found are interpreted by fail2ban in its own system time zone, and that may turn to be inappropriate. While the best practice is to configure the monitored applications to include explicit offsets, this option is meant to handle cases where that is not possible.

The supported time zones in this option are those with fixed offset: Z, UTC[+-]hhmm (you can also use GMT as an alias to UTC).

This option has no effect on log lines on which an explicit time zone has been found. Examples:

        logtimezone = UTC
        logtimezone = UTC+0200
        logtimezone = GMT-0100
banning action (default iptables-multiport) typically specified in the [DEFAULT] section for all jails.
This parameter will be used by the standard substitution of action and can be redefined central in the [DEFAULT] section inside jail.local (to apply it to all jails at once) or separately in each jail, where this substitution will be used.
the same as banaction but for some "allports" jails like "pam-generic" or "recidive" (default iptables-allports).
action(s) from /etc/fail2ban/action.d/ without the .conf/.local extension.
Arguments can be passed to actions to override the default values from the [Init] section in the action file. Arguments are specified by:

[name=value,name2=value,name3="values,values"]

Values can also be quoted (required when value includes a ","). More that one action can be specified (in separate lines).
boolean value (default true) indicates the banning of own IP addresses should be prevented
list of IPs not to ban. They can include a DNS resp. CIDR mask too. The option affects additionally to ignoreself (if true) and don't need to contain own DNS resp. IPs of the running host.
command that is executed to determine if the current candidate IP for banning (or failure-ID for raw IDs) should not be banned. This option operates alongside the ignoreself and ignoreip options. It is executed first, only if neither ignoreself nor ignoreip match the criteria.
IP will not be banned if command returns successfully (exit code 0). Like ACTION FILES, tags like <ip> are can be included in the ignorecommand value and will be substituted before execution.
provide cache parameters (default disabled) for ignore failure check (caching of the result from ignoreip, ignoreself and ignorecommand), syntax: 
        ignorecache = key="<F-USER>@<ip-host>", max-count=100, max-time=5m
        ignorecommand = if [ "<F-USER>" = "technical" ] && [ "<ip-host>" = "my-host.example.com" ]; then exit 0; fi;
                        exit 1
This will cache the result of ignorecommand (does not call it repeatedly) for 5 minutes (cache time) for maximal 100 entries (cache size), using values substituted like "user@host" as cache-keys. Set option ignorecache to empty value disables the cache.
effective ban duration (in seconds or time abbreviation format).
time interval (in seconds or time abbreviation format) before the current time where failures will count towards a ban.
number of failures that have to occur in the last findtime seconds to ban the IP.
backend to be used to detect changes in the logpath.
It defaults to "auto" which will try "pyinotify", "systemd" before "polling". Any of these can be specified. "pyinotify" is only valid on Linux systems with the "pyinotify" Python libraries.
use DNS to resolve HOST names that appear in the logs. By default it is "warn" which will resolve hostnames to IPs however it will also log a warning. If you are using DNS here you could be blocking the wrong IPs due to the asymmetric nature of reverse DNS (that the application used to write the domain name to log) compared to forward DNS that fail2ban uses to resolve this back to an IP (but not necessarily the same one). Ideally you should configure your applications to log a real IP. This can be set to "yes" to prevent warnings in the log or "no" to disable DNS resolution altogether (thus ignoring entries where hostname, not an IP is logged)..
regex (Python regular expression) to parse a common part containing in every message (see prefregex in section FILTER FILES for details).
regex (Python regular expression) to be added to the filter's failregexes (see failregex in section FILTER FILES for details). If this is useful for others using your application please share you regular expression with the fail2ban developers by reporting an issue (see REPORTING BUGS below).
regex which, if the log line matches, would cause Fail2Ban not consider that line. This line will be ignored even if it matches a failregex of the jail or any of its filters.
max number of matched log-lines the jail would hold in memory per ticket. By default it is the same value as maxretry of jail (or default). This option also affects values resolvable via tag <matches> in actions.

Available options are listed below.

requires pyinotify (a file alteration monitor) to be installed. If pyinotify is not installed, Fail2ban will use auto.
uses a polling algorithm which does not require external libraries.
uses systemd python library to access the systemd journal. Specifying logpath is not valid for this backend and instead utilises journalmatch from the jails associated filter config. Multiple systemd-specific flags can be passed to the backend, including journalpath and journalfiles, to explicitly set the path to a directory or set of files. journalflags, which by default is 4 and excludes user session files, can be set to include them with journalflags=1, see the python-systemd documentation for other settings and further details. Examples:
backend = systemd[journalpath=/run/log/journal/machine-1]
backend = systemd[journalfiles="/path/to/system.journal, /path/to/user.journal"]
backend = systemd[journalflags=1]

Each jail can be configured with only a single filter, but may have multiple actions. By default, the name of a action is the action filename, and in the case of Python actions, the ".py" file extension is stripped. Where multiple of the same action are to be used, the actname option can be assigned to the action to avoid duplication e.g.:

[ssh-iptables-ipset]
enabled = true
action = smtp.py[dest=chris@example.com, actname=smtp-chris]
         smtp.py[dest=sally@example.com, actname=smtp-sally]

The time entries in fail2ban configuration (like findtime or bantime) can be provided as integer in seconds or as string using special abbreviation format (e. g. 600 is the same as 10m).

years?, yea?, yy?
months?, mon?
weeks?, wee?, ww?
days?, da, dd?
hours?, hou?, hh?
minutes?, min?, mm?
seconds?, sec?, ss?
The question mark (?) means the optional character, so day as well as days can be used.

You can combine multiple tokens in format (separated with space resp. without separator), e. g.: 1y 6mo or 1d12h30m.
Note that tokens m as well as mm means minutes, for month use abbreviation mo or mon.

The time format can be tested using fail2ban-client:

fail2ban-client --str2sec 1d12h

Action files specify which commands are executed to ban and unban an IP address.

Like with jail.conf files, if you desire local changes create an [actionname].local file in the /etc/fail2ban/action.d directory and override the required settings.

Action files have two sections, Definition and Init .

The [Init] section enables action-specific settings. In jail.conf/jail.local these can be overridden for a particular jail as options of the action's specification in that jail.

The following commands can be present in the [Definition] section.

command(s) executed when the jail starts.
command(s) executed when the jail stops.
command(s) ran before any other action. It aims to verify if the environment is still ok.
command(s) that bans the IP address after maxretry log lines matches within last findtime seconds.
command(s) that unbans the IP address after bantime.

The [Init] section allows for action-specific settings. In jail.conf/jail.local these can be overwritten for a particular jail as options to the jail. The following are special tags which can be set in the [Init] section:

The maximum period of time in seconds that a command can executed, before being killed.

Commands specified in the [Definition] section are executed through a system shell so shell redirection and process control is allowed. The commands should return 0, otherwise error would be logged. Moreover if actioncheck exits with non-0 status, it is taken as indication that firewall status has changed and fail2ban needs to reinitialize itself (i.e. issue actionstop and actionstart commands). Tags are enclosed in <>. All the elements of [Init] are tags that are replaced in all action commands. Tags can be added by the fail2ban-client using the "set <JAIL> action <ACT>" command. <br> is a tag that is always a new line (\n).

More than a single command is allowed to be specified. Each command needs to be on a separate line and indented with whitespace(s) without blank lines. The following example defines two commands to be executed.


actionban = iptables -I fail2ban-<name> --source <ip> -j DROP
echo ip=<ip>, match=<match>, time=<time> >> /var/log/fail2ban.log

The following tags are substituted in the actionban, actionunban and actioncheck (when called before actionban/actionunban) commands.

IPv4 IP address to be banned. e.g. 192.168.0.2
number of times the failure occurred in the log file. e.g. 3
As per failures, but total of all failures for that ip address across all jails from the fail2ban persistent database. Therefore the database must be set for this tag to function.
As per ipfailures, but total based on the IPs failures for the current jail.
UNIX (epoch) time of the ban. e.g. 1357508484
concatenated string of the log file lines of the matches that generated the ban. Many characters interpreted by shell get escaped to prevent injection, nevertheless use with caution.
As per matches, but includes all lines for the IP which are contained with the fail2ban persistent database. Therefore the database must be set for this tag to function.
As per ipmatches, but matches are limited for the IP and for the current jail.

Python based actions can also be used, where the file name must be [actionname].py. The Python file must contain a variable Action which points to Python class. This class must implement a minimum interface as described by fail2ban.server.action.ActionBase, which can be inherited from to ease implementation.

Filter definitions are those in /etc/fail2ban/filter.d/*.conf and filter.d/*.local.

These are used to identify failed authentication attempts in log files and to extract the host IP address (or hostname if usedns is true).

Like action files, filter files are ini files. The main section is the [Definition] section.

There are several standard filter definitions used in the [Definition] section:

is the regex (regular expression) to parse a common part containing in every message, which is applied after datepattern found a match, before the search for any failregex or ignoreregex would start.
If this regex doesn't match the process is starting immediately with next message and search for any failregex does not occur.
If prefregex contains <F-CONTENT>...</F-CONTENT>, the part of message enclosed between this tags will be extracted and herafter used as whole message for search with failregex or ignoreregex.
For example: 
        prefregex = ^%(__prefix_line)s (?:ERROR|FAILURE) <F-CONTENT>.+</F-CONTENT>$
        failregex = ^user not found
                    ^authentication failed
                    ^unknown authentication method
You can use prefregex in order to:
- specify 1 common regex to match some common part present in every messages (do avoid unneeded match in every failregex if you have more as one);
- to cut some interesting part of message only (to simplify failregex) enclosed between tags <F-CONTENT> and </F-CONTENT>;
- to gather some failure identifier (e. g. some prefix matched by <F-MLFID>...<F-MLFID/> tag) to identify several messages belonging to same session, where a connect message containing IP followed by failure message(s) that are not contain IP; this provides a new multi-line parsing method as replacement for old (slow an ugly) multi-line parsing using buffering window (maxlines > 1 and <SKIPLINES>);
- to ignore some wrong, too long or even unneeded messages (a.k.a. parasite log traffic) which can be also present in journal, before failregex search would take place.
is the regex (regular expression) that will match failed attempts. The standard replacement tags can be used as part of the regex:
<HOST> - common regex for IP addresses and hostnames (if usedns is enabled). Fail2Ban will work out which one of these it actually is.
<ADDR> - regex for IP addresses (both families).
<IP4> - regex for IPv4 addresses.
<IP6> - regex for IPv6 addresses.
<DNS> - regex to match hostnames.
<CIDR> - helper regex to match CIDR (simple integer form of net-mask).
<SUBNET> - regex to match sub-net addresses (in form of IP/CIDR, also single IP is matched, so part /CIDR is optional).
<F-ID>...</F-ID> - free regex capturing group targeting identifier used for ban (instead of IP address or hostname).
<F-*>...</F-*> - free regex capturing named group stored in ticket, which can be used in action. 
For example <F-USER>[^@]+</F-USER> matches and stores a user name, that can be used in action with interpolation tag <F-USER>.
<F-ALT_*n>...</F-ALT_*n> - free regex capturing alternative named group stored in ticket.
For example first found matched value defined in regex as <F-ALT_USER>, <F-ALT_USER1> or <F-ALT_USER2> would be stored as <F-USER> (if direct match is not found or empty).
Every of abovementioned tags can be specified in prefregex and in failregex, thereby if specified in both, the value matched in failregex overwrites a value matched in prefregex.
NOTE: the failregex will be applied to the remaining part of message after prefregex processing (if specified), which in turn takes place after datepattern processing (whereby the string of timestamp matching the best pattern, cut out from the message).
For multiline regexs (parsing with maxlines greater that 1) the tag <SKIPLINES> can be used to separate lines. This allows lines between the matched lines to continue to be searched for other failures. The tag can be used multiple times.
This is an obsolete handling and if the lines contain some common identifier, better would be to use new handling (with tags <F-MLFID>...<F-MLFID/>).
is the regex to identify log entries that should be ignored by Fail2Ban, even if they match failregex.

specifies the maximum number of lines to buffer to match multi-line regexs. For some log formats this will not required to be changed. Other logs may require to increase this value if a particular log file is frequently written to.
specifies a custom date pattern/regex as an alternative to the default date detectors e.g. %%Y-%%m-%%d %%H:%%M(?::%%S)?.
For a list of valid format directives, see Python library documentation for strptime behaviour.
NOTE: due to config file string substitution, that %'s must be escaped by an % in config files.
Also, special values of Epoch (UNIX Timestamp), TAI64N and ISO8601 can be used as datepattern.
Normally the regexp generated for datepattern additionally gets word-start and word-end boundaries to avoid accidental match inside of some word in a message.
There are several prefixes and words with special meaning that could be specified with custom datepattern to control resulting regex:
{DEFAULT} - can be used to add default date patterns of fail2ban.
{DATE} - can be used as part of regex that will be replaced with default date patterns.
{^LN-BEG} - prefix (similar to ^) changing word-start boundary to line-start boundary (ignoring up to 2 characters). If used as value (not as a prefix), it will also set all default date patterns (similar to {DEFAULT}), but anchored at begin of message line.
{UNB} - prefix to disable automatic word boundaries in regex.
{NONE} - value would allow one to find failures totally without date-time in log message. Filter will use now as a timestamp (or last known timestamp from previous line with timestamp).
specifies the systemd journal match used to filter the journal entries. See journalctl(1) and systemd.journal-fields(7) for matches syntax and more details on special journal fields. This option is only valid for the systemd backend.
Similar to actions, filters may have an [Init] section also (optional since v.0.10). All parameters of both sections [Definition] and [Init] can be overridden (redefined or extended) in jail.conf or jail.local (or in related filter.d/filter-name.local).
Every option supplied in the jail to the filter overwrites the value specified in [Init] section, which in turm would overwrite the value in [Definition] section.
Besides the standard settings of filter both sections can be used to initialize filter-specific options.
Filters can also have a section called [INCLUDES]. This is used to read other configuration files.
before
indicates that this file is read before the [Definition] section.
after
indicates that this file is read after the [Definition] section.

Fail2ban was originally written by Cyril Jaquier <cyril.jaquier@fail2ban.org>. At the moment it is maintained and further developed by Yaroslav O. Halchenko <debian@onerussian.com>, Daniel Black <daniel.subs@internode.on.net> and Steven Hiscocks <steven-fail2ban@hiscocks.me.uk> along with a number of contributors. See THANKS file shipped with Fail2Ban for a full list. Manual page written by Daniel Black and Yaroslav Halchenko.

Report bugs to https://github.com/fail2ban/fail2ban/issues

Copyright © 2013 the Fail2Ban Team
Copyright of modifications held by their respective authors.
Licensed under the GNU General Public License v2 (GPL) or (at your option) any later version.

fail2ban-server(1)

November 2015 Fail2Ban