Any audit trail file not named on the command line must conform to the audit trail filename format. Files produced by the audit system already have this format. Output file names produced by auditreduce are in this format. It is:

start-time.end-time.suffix

where start-time is the 14-character timestamp of when the file was opened, end-time is the 14-character timestamp of when the file was closed, and suffix is the name of the machine which generated the audit trail file, or some other meaningful suffix (e.g., all, if the file contains a combined group of records from many machines). The end-time may be the literal string not_terminated, to indicate that the file is still being written to by the audit system. Timestamps are of the form yyyymmddhhmmss (year, month, day, hour, minute, second). The timestamps are in Greenwich Mean Time (GMT).

The file selection options indicate which files are to be processed and certain types of special treatment.

-A

All of the records from the input files will be selected regardless of their timestamp. This option effectively disables the -a, -b, and -d options. This is useful in preventing the loss of records if the -D option is used to delete the input files after they are processed. Note, however, that if a record is not selected due to another option, then -A will not override that.

-C

Only process complete files. Files whose filename end-time timestamp is not_terminated are not processed (such a file is currently being written to by the audit system). This is useful in preventing the loss of records if -D is used to delete the input files after they are processed. It does not apply to files specified on the command line.

-D suffix

Delete input files after they deleted if the entire run is successful. If auditreduce detects an error while reading a file, then that file is not deleted. If -D is specified, -A, -C and -O are also implied. suffix is given to the -O option. This helps prevent the loss of audit records by ensuring that all of the records are written, only complete files are processed, and the records are written to a file before being deleted. Note that if both -D and -O are specified in the command line, the order of specification is significant. The suffix associated with the latter specification is in effect.

-M machine

Allows selection of records from files with machine as the filename suffix. If -M is not specified, all files are processed regardless of suffix. -M can also be used to allow selection of records from files that contain combined records from many machines and have a common suffix (such as all).

-N

Select objects in new mode.This flag is off by default, thus retaining backward compatibility. In the existing, old mode, specifying the -e, -f, -g, -r, or -u flags would select not only actions taken with those IDs, but also certain objects owned by those IDs. When running in new mode, only actions are selected. In order to select objects, the -o option must be used.

-O suffix

Direct output stream to a file in the currenti audit_root_dir with the indicated suffix. suffix may alternatively contain a full pathname, in which case the last component is taken as the suffix, ahead of which the timestamps will be placed, ahead of which the remainder of the pathname will be placed. If the -O option is not specified, the output is sent to the standard output. When auditreduce places timestamps in the filename, it uses the times of the first and last records in the merge as the start-time and end-time.

-Q

Quiet. Suppress notification about errors with input files.

-R pathname

Specify the pathname of an alternate audit root directory audit_root_dir to be pathname. Therefore, rather than using /etc/security/audit/*/files by default, pathname/*/files will be examined instead.

-S server

This option causes auditreduce to read audit trail files from a specific location (server directory). server is normally interpreted as the name of a subdirectory of the audit root, therefore auditreduce will look in audit_root_dir/server/files for the audit trail files. But if server contains any `/' characters, it is the name of a specific directory not necessarily contained in the audit root. In this case, server/files will be consulted. This option allows archived files to be manipulated easily, without requiring that they be physically located in a directory structure like that of /etc/security/audit.

-V

Verbose. Display the name of each file as it is opened, and how many records total were written to the output stream.

The record selection options listed below are used to indicate which records are written to the output file produced by auditreduce.

Multiple arguments of the same type are not permitted.

-a date-time

Select records that occurred at or after date-time. The date-time argument is described under Option Arguments, below. date-time is in local time. The -a and -b options can be used together to form a range.

-b date-time

Select records that occurred before date-time.

-c audit-classes

Select records by audit class. Records with events that are mapped to the audit classes specified by audit-classes are selected. Audit class names are defined in audit_class(4). The audit-classes can be a comma separated list of audit flags like those described in audit_control(4). Using the audit flags, one can select records based upon success and failure criteria.

-d date-time

Select records that occurred on a specific day (a 24-hour period beginning at 00:00:00 of the day specified and ending at 23:59:59). The day specified is in local time. The time portion of the argument, if supplied, is ignored. Any records with timestamps during that day are selected. If any hours, minutes, or seconds are given in time, they are ignored. -d can not be used with -a or -b.

-e effective-user

Select records with the specified effective-user.

-f effective-group

Select records with the specified effective-group.

-g real-group

Select records with the specified real-group.

-j subject-ID

Select records with the specified subject-ID where subject-ID is a process ID.

-m event

Select records with the indicated event. The event is the literal string or the event number.

-o object_type=objectID_value

Select records by object type. A match occurs when the record contains the information describing the specified object_type and the object ID equals the value specified by objectID_value. The allowable object types and values are as follows:

file=pathname

Select records containing file system objects with the specified pathname, where pathname is a comma separated list of regular expressions. If a regular expression is preceeded by a tilde (~), files matching the expression are excluded from the output. For example, the option file=~/usr/openwin,/usr,/etc would select all files in /usr or /etc except those in /usr/openwin. The order of the regular expressions is important because auditreduce processes them from left to right, and stops when a file is known to be eitherselected or excluded. Thus the option file= /usr, /etc, ~/usr/openwin would select all files in /usr and all files in /etc. Files in /usr/openwin are not excluded because the regular expression /usr is matched first. Care should be given in surrounding the pathname with quotes so as to prevent the shell from expanding any tildes.

filegroup=group

Select records containing file system objects with group as the owning group.

fileowner=user

Select records containing file system objects with user as the owning user.

msgqid=ID

Select records containing message queue objects with the specified ID where ID is a message queue ID.

msgqgroup=group

Select records containing message queue objects with group as the owning or creating group.

msgqowner=user

Select records containing message queue objects with user as the owning or creating user.

pid=ID

Select records containing process objects with the specified ID where ID is a process ID. Process are objects when they are receivers of signals.

procgroup=group

Select records containing process objects with group as the real or effective group.

procowner=user

Select records containing process objects with user as the real or effective user.

semid=ID

Select records containing semaphore objects with the specified ID where ID is a semaphore ID.

semgroup=group

Select records containing semaphore objects with group as the owning or creating group.

semowner=user

Select records containing semaphore objects with user as the owning or creating user.

shmid=ID

Select records containing shared memory objects with the specified ID where ID is a shared memory ID.

shmgroup=group

Select records containing shared memory objects with group as the owning or creating group.

shmowner=user

Select records containing shared memory objects with user as the owning or creating user.

sock=port_number|machine

Select records containing socket objects with the specified port_number or the specified machine where machine is a machine name as defined in hosts(4).

-r real-user

Select records with the specified real-user.

-u audit-user

Select records with the specified audit-user. When one or more filename arguments appear on the command line, only the named files are processed. Files specified in this way need not conform to the audit trail filename format. However, -M, -S, and -R may not be used when processing named files. If the filename is ``-'' then the input is taken from the standard input.

audit-trail-file

An audit trail file as defined in audit.log(4). An audit trail file not named on the command line must conform to the audit trail file name format. Audit trail files produced as output of auditreduce are in this format as well. The format is:

start-time . end-time . suffix

start-time is the 14 character time stamp denoting when the file was opened. end-time is the 14 character time stamp denoting when the file was closed. end-time may also be the literal string not_terminated, indicating the file is still be written to by the audit daemon or the file was not closed properly (a system crash or abrupt halt occurred). suffix is the name of the machine that generated the audit trail file (or some other meaningful suffix; e.g. all would be a good suffix if the audit trail file contains a combined group of records from many machines).

date-time

The date-time argument to -a, -b, and -d can be of two forms: An absolute date-time takes the form:

yyyymmdd [ hh [ mm [ ss ]]]

where yyyy specifies a year (with 1970 as the earliest value), mm is the month (01-12), dd is the day (01-31), hh is the hour (00-23), mm is the minute (00-59), and ss is the second (00-59). The default is 00 for hh, mm and ss.

An offset can be specified as: +n d|h|m| s where n is a number of units, and the tags d, h, m, and s stand for days, hours, minutes and seconds, respectively. An offset is relative to the starting time. Thus, this form can only be used with the -b option.

event

The literal string or ordinal event number as found in audit_event(4). If event is not found in the audit_event file it is considered invalid.

group

The literal string or ordinal group ID number as found in group(4). If group is not found in the group file it is considered invalid. group may be negative.

pathname

A regular expression describing a pathname.

user

The literal username or ordinal user ID number as found in passwd(4). If the username is not found in the passwd file it is considered invalid. user may be negative.

praudit(1M) is available to display audit records in a human-readable form.

This will display the entire audit trail in a human-readable form:

% auditreduce | praudit

If all the audit trail files are being combined into one large file, then deleting the original files could be desirable to prevent the records from appearing twice:

% auditreduce -V -d /etc/security/audit/combined/all

This will print what user milner did on April 13, 1988. The output will be displayed in a human-readable form to the standard output:

% auditreduce -d 19880413 -u milner | praudit

The above example may produce a large volume of data if milner has been busy. Perhaps looking at only login and logout times would be simpler. The -c option will select records from a specified class:

% auditreduce -d 19880413 -u milner -c lo | praudit

To see milner's login/logout activity for April 13, 14, and 15 the following is used. The results are saved to a file in the current working directory. Note that the name of the output file will have milnerlo as the suffix, with the appropriate timestamp prefixes. Note that the long form of the name is used for the -c option:

% auditreduce -a 19880413 -b +3d -u milner -c login_logout -o milnerlo

To follow milner's movement about the file system on April 13, 14, and 15 the chdir record types could be viewed. Note that in order to get the same time range as the above example we needed to specify the -b time as the day after our range. This is because 19880416 defaults to midnight of that day, and records before that fall on 0415, the end-day of the range.

% auditreduce -a 19880413 -b 19880416 -u milner -m AUE_CHDIR | praudit

In this example the audit records are being collected in summary form (the login/logout records only). The records are being written to a summary file in a different directory than the normal audit root to prevent the selected records from existing twice in the audit root.

% auditreduce -d 19880330 -c lo -o /etc/security/audit_summary/logins

If activity for user ID 9944 has been observed, but that user is not known to the system administrator, then the following example will search the entire audit trail for any records generated by that user. auditreduce will query the system as to the current validity of ID 9944, and print a warning message if it is not currently active:

% auditreduce -o /etc/security/audit_suspect/user9944 -u 9944