Novell SUSE Linux Enterprise 11 Quick Start

Linux Audit Quick Start
SUSE Linux Enterprise 11 SP1
Linux audit allows you to comprehensively log and track any access to les, directories, or resources of your system and trace system calls. It enables you to monitor your system for application misbehavior or code mal­functions. By creating a sophisticated set of rules including le watches and system call auditing, you can make sure that any violation of your security policies is noticed and properly addressed.
NOVELL® QUICK START CARD
To set up Linux audit on your system, proceed as follows:
1. Stop the audit daemon that is running by default with the rcauditd stop command.
2. Adjust the system conguration for audit and enable audit.
3. Congure the audit daemon.
4. Determine which system components to audit and set up audit rules.
5. Start the audit daemon after you have completed the conguration of the audit system using the rcauditd start command.
6. Determine which reports to run and congure these reports.
7. Analyze the audit logs and reports.
8. (Optional) Analyze individual system calls with autrace.
IMPORTANT: Users Entitled to Work with Audit
The audit tools, conguration les, and logs are only available to root. This protects audit from ordinary users of the system. To manipulate any aspect of audit, you must be logged in as root.

Enabling Audit

Your rst task enabling audit is to activate system call audit­ing, since system call auditing capabilities are needed even when you are only conguring plain le or directory watches:
Enabling System Call Auditing for One Session Only
Enable with auditctl -e 1 and disable with au- ditctl -e 0. These settings are not persistent and do not survive a reboot.
Enabling System Call Auditing Permanently
Permanently enable audit contexts for system calls by changing AUDITD_DISABLE_CONTEXTS in /etc/ sysconfig/auditd from yes to no. To permanently disable audit contexts for system calls, revert this setting to yes. THis conguration will be applied with the next start of the audit daemon.
Conguring Audit
The conguration of the audit daemon is contained in the /etc/audit/auditd.conf conguration le. The default settings as shipped with SUSE Linux Enterprise should be sufcient for most setups.
log_file = /var/log/audit/audit.log log_format = RAW log_group = root priority_boost = 4
1
flush = INCREMENTAL freq = 20 num_logs = 4 disp_qos = lossy dispatcher = /usr/sbin/audispd name_format = NONE #name = mydomain max_log_file = 5 max_log_file_action = ROTATE space_left = 75 space_left_action = SYSLOG action_mail_acct = root admin_space_left = 50 admin_space_left_action = SUSPEND disk_full_action = SUSPEND disk_error_action = SUSPEND #tcp_listen_port = tcp_listen_queue = 5 #tcp_client_ports = 1024-65535 tcp_client_max_idle = 0
Most of the settings in this le concern the audit log les and how the logging is done. The most important settings all concern the actions the daemon should take when en­countering certain critical conditions or errors (system low on disk space, system out of disk space, or disk error) and when to warn the administrator about these conditions. These actions are customizable and range from a mere warning in syslog to a complete halt of the system. For more information about /etc/audit/auditd.conf, refer to The Linux Audit Framework manual and the manual page of auditd.conf (auditd.conf(8)).

Setting Up Audit Rules

Audit rules are used to specify which components of your system are audited. There are three basic types of audit rules:
• Basic audit system parameters
• File and directory watches
• System call audits
Before creating an audit rule set and before rolling it out to your system, carefully determine which components to audit. Extensive auditing causes a substantial logging load. Make sure that your system provides enough disk space to store large audit logs and test your audit rule set extensively before rolling it out to a production system.
Audit rules can either be passed to the audit system by the command line using auditctl or bundled into a rules le located under /etc/audit/audit.rules that is read during the start of the audit daemon:
# basic audit system parameters
-D
-b 8192
-f 1
-e 1
# some file and directory watches
-w /var/log/audit/
-w /etc/audit/auditd.conf -p rxwa
-w /etc/audit/audit.rules -p rxwa
-w /etc/passwd -p rwxa
-w /etc/sysconfig/
# an example system call rule
-a entry,always -S umask
The basic audit system parameters include a rule to delete any preexisting rules (-D) to avoid clashes with the new rules, a rule that sets the number of outstanding audit buffers (-b), the failure ag (-f), and the enable ag (-e):
-b
Depending on the audit load of your system, increase or decrease the number of outstanding audit buffers. If there are no more buffers left, the kernel checks the failure ag for action.
-f
The failure ag controls the kernel's reaction to critical errors. Possible values are 0 (silent), 1 (printk, print a failure message), and 2 (panic, bring the system down—no clean shutdown and risk of data loss or cor­ruption).
-e
If set to 1, this enables audit and audit contexts for sys­tem calls. Setting it to 2 does the same, but also locks down the conguration. Set to 0, audit is disabled. This ag is used to enable or disable audit temporarily.
File system watches can be added whenever you want to track les or directories for unauthorized access. Typical examples would include watching the audit conguration and logs and user and security databases. Use permission ltering to focus on those system calls requesting the per­missions in which you are interested:
-w /etc/audit/audit.rules -p rxwa
The -p ag enables permission ltering. This example has permission ltering turned on for read, write, execute, and attribute change permissions.
Note the following limitations to le system watches:
• Directory watches produce less verbose logs than exact le watches. When in need of detailed le-related records, enable separate le watches for all les of inter­est.
2
• Pathname globbing of any kind is not supported by audit. Always use the exact pathnames.
• Auditing can only be performed on existing les. Any les added while the audit daemon is already running are ig­nored until the audit rule set is updated to watch the new les.
Assigning keys to your audit rules helps you to identify any records related to this rule in the logs. An example rule plus key:
-w /etc/var/log/audit/ -k LOG_audit
The -k option attaches a text string to any event that is recorded in the logs due to this rule. Using the ausearch log analyzer, you can easily lter for any events related to this particular rule.
A sample system call audit rule could look like the following:
-a entry,always -S umask
This adds the rule to the system call entry list (-a) and logs an event whenever this system call is used (entry,al-
ways). The -S option precedes the actual system call, umask in this example. Using -F, you could add optional
ltering to this rule. For more information about audit rules, refer to The Linux Audit Framework and the manual page of auditctl (auditctl(8)).

Generating Reports

Every audit event is recorded in the audit log, /var/log/ audit/audit.log. To avoid having to read the raw audit
log, congure custom audit reports with aureport and run them regularly. Use the aureport tool to create various types of reports ltering for different elds of the audit records in the log. The output of any aureport command is printed in column format and can easily be piped to other commands for further processing. Because the aureport commands are scriptable, you can easily create custom re­port scripts to run at certain intervals to gather the audit information for you.
aureport --failed
Run this report to get statistics of failed events on your system. This report includes the same event categories as the summary report. To get detailed information for a particular event type, run the individual report adding the --failed option to lter for failed events of this type, such as aureport -f --failed to display all failed le-related events.
aureport -l
Run this command to generate a numbered list of all login-related events. The report includes date, time, audit ID, host and terminal used, name of the executable, success or failure of the attempt, and an event ID.
aureport -p
Run this report to generate a numbered list of all pro­cess-related events. This command generates a numbered list of all process events including date, time, process ID, name of the executable, system call, audit ID, and event number.
aureport -f
Run this report to generate a numbered list of all le­related events. This command generates a numbered list of all process events including date, time, process ID, name of the executable, system call, audit ID and event number.
aureport -u
Run this report to nd out which users are running what executables on your system. This command generates a numbered list of all user-related events including date, time, audit ID, terminal used, host, name of the exe­cutable, and an event ID.
Use the -ts and -te (for start time and end time) options with any of the above commands to limit your reports to a certain time frame. Use the -i option with any of these commands to transform numeric entities to human-readable text. The following command creates a le report for the time between 8 am and 5:30 pm on the current day and converts numeric entries to text.
aureport -ts 8:00 -te 17:30 -f -i
aureport --summary
Run this report to get a rough overview of the current audit statistics (events, logins, processes, etc.). To get detailed information about any of the event categories listed, run individual reports for the event type.
aureport --success
Run this report to get statistics of successful events on your system. This report includes the same event cate­gories as the summary report. To get detailed informa­tion for a particular event type, run the individual report adding the --success option to lter for successful events of this type, for example, aureport -f -- success to display all successful le-related events.

Analyzing Audit Log Files and Reports

While aureport helps you generate custom reports focusing on a certain area, ausearch helps you to nd the detailed log entry of individual events:
ausearch -a audit_event_id
Run this search to view all records carrying a particular audit event ID. Each audit event message is logged along with a message ID consisting of a UNIX epoch time stamp plus a unique event ID separated by a colon. All events that are logged from one application's system call have the same event ID. For example, use ausearch -a 1234 to display all audit events carrying this audit event ID. As one application's system call may trigger several
3
events to be logged, you are likely to retrieve more than one record from the log.
ausearch -ul login_id
Run this search to view records associated with a partic­ular login user ID. It displays any records related to the user login ID specied provided that user had been able to log in successfully. For example, use ausearch -ul root to list all processes owned by the given login user ID.
ausearch -k key
Run this search to nd records that contain a certain key assigned in the audit rule set. For example, use ause- arch -k CFG_etc to display any records containing the CFG_etc key.
ausearch -m message_type
Run this search to nd records related to a particular message type. Examples of valid message types include
PATH, SYSCALL, USER_LOGIN. Invoking ausearch
-m without a message type displays a list of all message
types.
ausearch -f filename
Run this search to nd records containing a certain le­name. For example, run ausearch -f /foo/bar for all records related to the /foo/bar le. Using the le­name alone would work as well, but using relative paths would not.
ausearch -p process_id
Run this to search for records related to a certain process ID. For example, use ausearch -p 13368 to search for all records related to this process ID.
autrace /usr/bin/less /etc/sysconfig/auditd
Waiting to execute: /usr/bin/less Cleaning up... No rules Trace complete. You can locate the records with 'ausearch -i -p 7642'
Always use the full path to the executable to autrace. After the trace is complete, autrace provides you with the event ID of the trace, so you can analyze the entire data trail with ausearch. To restore the audit system to use the audit rule set again, just restart the audit daemon by calling rcauditd restart.

Audit Tool Set

auditctl
Controls the audit system. Check the audit daemon's status and rule set, delete rules, or create new ones.
aureport
Create various types of reports from the audit daemon logs.
ausearch
Create custom queries to search the audit daemon logs.
autrace
Add audit rules to trace a process. Similar to strace.
rcauditd
Controls the audit init script.

Files

Use the -ts and -te (for start time and end time) options with any of these commands to limit your reports to a cer­tain time frame. Use the -i option with any of these to transform numeric entities to human readable text. The following command searches for any le event related to audit.log that took place any time between 8 am and 5:30 pm on the current day and converts numeric entries to text.
ausearch -ts 8:00 -te 17:30 -f audit.log -i

Analyzing Individual System Calls

Perform dedicated audits of individual processes using the autrace command. autrace works similarly to the strace command, but gathers slightly different information. The output of autrace is written to /var/log/audit/audit .log and does not look any different from the standard audit log entries.
When performing an autrace on a process, make sure that any audit rules are purged from the queue to avoid having these rules clash with the ones autrace adds itself. Delete the audit rules with the auditctl -D command.
/etc/audit/auditd.conf
Contains conguration options specic to the audit daemon, such as log le location, log rotation, maximum size of the log le, and various actions to take when the system starts to run low on disk space.
/etc/sysconfig/auditd
Controls conguration aspects of auditd that are not covered in /etc/audit/auditd.conf, such as the locale to use with audit, the use of audit contexts with system calls, and whether rules and watches should be deleted on shutdown of the system.
/etc/audit/audit.rules
Controls the rules auditd processes to track system calls and le and directory access.
/var/log/audit/audit.log
The audit log le.

For More Information

For a more detailed introduction to the Linux audit frame­work, refer to the The Linux Audit Framework manual that is available at http://www.novell.com/
documentation/sles11/http://www.novell.com/ documentation/sled11/.
4

Legal Notice

All content is copyright © 2006- 2009 Novell, Inc.
distributed for noncommercial use only. The express autho­rization of Novell, Inc must be obtained prior to any other use of any manual or part thereof.
This manual is protected under Novell intellectual property rights. By reproducing, duplicating or distributing this manual you explicitly agree to conform to the terms and conditions of this license agreement.
This manual may be freely reproduced, duplicated and dis­tributed either as such or as part of a bundled package in electronic and/or printed format, provided however that the following conditions are fullled:
That this copyright notice and the names of authors and contributors appear clearly and distinctively on all repro­duced, duplicated and distributed copies. That this manual, specically for the printed format, is reproduced and/or
For Novell trademarks, see the Novell Trademark and Ser­vice Mark list http://www.novell.com/company/ legal/trademarks/tmlist.html. Linux* is a regis­tered trademark of Linus Torvalds. All other third party trademarks are the property of their respective owners. A trademark symbol (®, ™ etc.) denotes a Novell trademark; an asterisk (*) denotes a third party trademark.
All information found in this book has been compiled with utmost attention to detail. However, this does not guarantee complete accuracy. Neither Novell, Inc., SUSE LINUX Prod­ucts GmbH, the authors, nor the translators shall be held liable for possible errors or the consequences thereof.
5
Created by SUSE® with XSL-FO
6
Loading...