The syslog.conf file is the configuration file for the syslogd(8) program. It consists of blocks of lines separated by program and hostname specifications (separations appear along on the line), with each line containing two fields: the selector field which specifies the types of messages and priorities to which the line applies, and an action field which specifies the action to be taken if a message syslogd(8) receives matches the selection criteria. The selector field is separated from the action field by one or more tab characters or spaces.
Note that if you use spaces as separators, your syslog.conf might be incompatible with other Unices or Unix-like systems. This functionality was added for ease of configuration (e.g. it is possible to cut-and-paste into syslog.conf), and to avoid possible mistakes. This change however preserves backwards compatibility with the old style of syslog.conf (i.e., tab characters only).
The selectors are encoded as a facility, a period ("."), an optional set of comparison flags ([![<=>]]), and a level, with no intervening white-space. Both the facility and the level are case insensitive.
The facility describes the part of the system generating the message, and is one of the following keywords: auth, authpriv, console, cron, daemon, ftp, kern, lpr, mail, mark, news, ntp, security, syslog, user, uucp and local0 through local7. These keywords (with the exception of mark) correspond to similar "LOG_" values specified to the openlog(3) and syslog(3) library routines.
The comparison flags may be used to specify exactly what is logged. The default comparison is "=>" (or, if you prefer, ">="), which means that messages from the specified facility list, and of a priority level equal to or greater than level will be logged. Comparison flags beginning with "!" will have their logical sense inverted. Thus "!=info" means all levels except info and "!notice" has the same meaning as "<notice".
The level describes the severity of the message, and is a keyword from the following ordered list (higher to lower): emerg, alert, crit, err, warning, notice, info and debug. These keywords correspond to similar "LOG_" values specified to the syslog(3) library routine.
Each block of lines is separated from the previous block by a program or hostname specification. A block will only log messages corresponding to the most recent program and hostname specifications given. Thus, with a block which selects ppp as the program, directly followed by a block that selects messages from the hostname dialhost, the second block will only log messages from the ppp(8) program on dialhost.
A program specification is a line beginning with #!prog or !prog (the former is for compatibility with the previous syslogd, if one is sharing syslog.conf files, for example) and the following blocks will be associated with calls to syslog(3) from that specific program. A program specification for foo will also match any message logged by the kernel with the prefix "foo: ". The #!+prog or !+prog specification works just like the previous one, and the #!-prog or !-prog specification will match any message but the ones from that program. Multiple programs may be listed, separated by commas: !prog1,prog2 matches messages from either program, while !-prog1,prog2 matches all messages but those from prog1 or prog2.
A hostname specification of the form #+hostname or +hostname means the following blocks will be applied to messages received from the specified hostname. Alternatively, the hostname specification #-hostname or -hostname causes the following blocks to be applied to messages from any host but the one specified. If the hostname is given as @, the local hostname will be used. As for program specifications, multiple comma-separated values may be specified for hostname specifications.
A program or hostname specification may be reset by giving the program or hostname as *.
See syslog(3) for further descriptions of both the facility and level keywords and their significance. It is preferred that selections be made on facility rather than program, since the latter can easily vary in a networked environment. In some cases, though, an appropriate facility simply does not exist.
If a received message matches the specified facility and is of the specified level (or a higher level), and the first word in the message after the date matches the program, the action specified in the action field will be taken.
Multiple selectors may be specified for a single action by separating them with semicolon (";") characters. It is important to note, however, that each selector can modify the ones preceding it.
Multiple facilities may be specified for a single level by separating them with comma (",") characters.
An asterisk ("*") can be used to specify all facilities, all levels, or all programs.
The special facility "mark" receives a message at priority "info" every 20 minutes (see syslogd(8)). This is not enabled by a facility field containing an asterisk.
The special level "none" disables a particular facility.
The action field of each line specifies the action to be taken when the selector field selects a message. There are five forms:
- A pathname (beginning with a leading slash). Selected messages are appended to the file.
To ensure that kernel messages are written to disk promptly, syslog.conf calls fsync(2) after writing messages from the kernel. Other messages are not synced explicitly. You may prefix a pathname with the minus sign, "-", to forego syncing the specified file after every kernel message. Note that you might lose information if the system crashes immediately following a write attempt. Nevertheless, using the "-" option may improve performance, especially if the kernel is logging many messages.
- A hostname (preceded by an at ("@") sign). Selected messages are forwarded to the syslogd(8) program on the named host.
- A comma separated list of users. Selected messages are written to those users if they are logged in.
- An asterisk. Selected messages are written to all logged-in users.
- A vertical bar ("|"), followed by a command to pipe the selected messages to. The command is passed to sh(1) for evaluation, so usual shell metacharacters or input/output redirection can occur. (Note however that redirecting stdio(3) buffered output from the invoked command can cause additional delays, or even lost output data in case a logging subprocess exited with a signal.) The command itself runs with stdout and stderr redirected to /dev/null. Upon receipt of a SIGHUP, syslogd(8) will close the pipe to the process. If the process did not exit voluntarily, it will be sent a SIGTERM signal after a grace period of up to 60 seconds.
The command will only be started once data arrives that should be piped to it. If it exited later, it will be restarted as necessary. So if it is desired that the subprocess should get exactly one line of input only (which can be very resource-consuming if there are a lot of messages flowing quickly), this can be achieved by exiting after just one line of input. If necessary, a script wrapper can be written to this effect.
Unless the command is a full pipeline, it is probably useful to start the command with exec so that the invoking shell process does not wait for the command to complete. Warning: the process is started under the UID invoking syslogd(8), normally the superuser.
Blank lines and lines whose first non-blank character is a hash ("#") character are ignored.
The "kern" facility is usually reserved for messages generated by the local kernel. Other messages logged with facility "kern" are usually translated to facility "user". This translation can be disabled; see syslogd(8) for details.