From Gentoo Wiki
Jump to: navigation, search
Other languages:
English • ‎italiano • ‎日本語 • ‎한국어 • ‎русский


AIDE stands for Advanced Intrusion Detection Environment and is an application that scans files and other resources and stores information about these files in a database. This information can be hash information, file size, ownership and more. The application can then, once this database is available, rescan the system and compare the results with the previously stored values. If values differ, then the file is changed and this change is reported.

Installation and configuration

Within Gentoo, it is easy to install aide after setting the USE flags accordingly. The supported USE flags at the time of writing are:

USE flag (what is that?) Default Recommended Description
acl No Add support for Access Control Lists
audit No Yes Enable support for sys-process/audit
curl No Add support for client-side URL transfer library
mhash No Yes Add support for the mhash library
nls No Yes Add Native Language Support (using gettext - GNU locale utilities)
postgres No Add support for the postgresql database
prelink No Enable support for sys-devel/prelink
selinux No  !!internal use only!! Security Enhanced Linux support, this must be set by the selinux profile or breakage will occur
static No  !!do not set this during bootstrap!! Causes binaries to be statically linked instead of dynamically
xattr No Yes Add support for extended attributes (filesystem-stored metadata)
zlib No Yes Add support for zlib (de)compression

Then it is a matter of installing the software:

root #emerge --ask app-forensics/aide

The configuration file for aide is not as daunting as it might seem at first sight. The default file is stored at /etc/aide/aide.conf but administrators can easily create multiple separate configuration files if necessary. Besides a few variables, the configuration file contains a few short-hand notations for what aspects of files to scan for (only hashes, or also inode information, etc.) and then which files to scan.

Let's first look at the variables.

FILE aide.conf variables

These parameters define where the database is stored that contains the known values (database) and where to store a new database when a new one (database_out) is created. It is generally recommended to not have these variables point to the same, instead manually copying over the generated database from one location to the other.

For now, leave those variables as-is, we'll get back to them later.

FILE aide.conf shorthand notations
Binlib = p+i+n+u+g+s+b+m+c+md5+sha1
Logs = p+i+n+u+g+S

These are short-hand notations for what to measure. The letters are described in the default aide.conf file, but the next table gives an overview of the most common ones.

Short Description
p Permissions
i inode number
n Number of (hard)links
u User information
g Group information
s Size
S Size (only report when the size is suddenly smaller - growing is allowed)
b Block count
m Modification time

Also, it is pretty obvious that md5 and sha1 mean that the MD5 and SHA-1 checksums are taken respectively.

These short-hand notations are then used to identify what to scan for which files.

FILE aide.conf scan targets
/bin Binlib
/sbin Binlib
/var/log Logs

This is the overview of which directories to scan, and what to scan for. In the above three lines example, we tell AIDE to scan the /bin and /sbin locations and take the measures identified earlier in the Binlib short-hand notation. The /var/log location should use the Logs scan measures.

AIDE supports regular expressions and users are allowed to "remove" matches. For instance, to scan /var/log but not /var/log/portage then include an exclusion set like so:

FILE aide.conf other scan targets
/var/log Logs

Initialization and frequent scanning

First we need to initialize the database once.

root #aide --init --config=/etc/aide/aide.conf
AIDE, version 0.14.2
### AIDE database at /var/lib/aide/ initialized.

Once initialized, we can copy over the database file.

root #cd /var/lib/aide; cp aide.db

With the database now available, we can scan the entries again for potential modifications:

root #aide --check --config=/etc/aide/aide.conf
AIDE, version 0.14.2
### All files match AIDE database. Looks okay!

When a file modification occurred, a notification will be sent out:

root #aide --check --config=/etc/aide/aide.conf
AIDE found differences between database and filesystem!!
Start timestamp: 2013-04-11 15:31:02
  Total number of files:        318
  Added files:                  0
  Removed files:                0
  Changed files:                2
Changed files:
changed: /etc/pam.d
changed: /etc/pam.d/run_init
Detailed information about changes:
Directory: /etc/pam.d
  Mtime    : 2013-04-09 22:11:18              , 2013-04-11 15:31:01
  Ctime    : 2013-04-09 22:11:18              , 2013-04-11 15:31:01
File: /etc/pam.d/run_init
  Size     : 205                              , 208
  Mtime    : 2013-04-09 22:11:18              , 2013-04-11 15:31:00
  Ctime    : 2013-04-09 22:11:18              , 2013-04-11 15:31:01
  Inode    : 394203                           , 394053
  MD5      : Mm0KPzpPt63eqGClTJ/KaQ==         , eLUrP2BsIq25f3AZX+dlBA==
  SHA1     : NrQtsUeOsXS4RHUq+ejYBne5V6E=     , 5A6ef6VJCcMiqEjKQ7e9xkBNZB8=

Best Practices

Be clear with what to scan

The default AIDE configuration is useful, but it needs to be fine-tuned to suit the users' needs. It is important to know which files to scan and why.

For instance, to scan for all authentication-related files but not for other files, use a configuration like so:

FILE aide.conf authentication-related scan targets
# SELinux policy and settings
/etc/selinux ConfFiles
# Authentication databases
/etc/passwd ConfFiles
/etc/shadow ConfFiles
/etc/nsswitch.conf ConfFiles
# Authentication configuration
/etc/pam.d ConfFiles
/etc/securetty ConfFiles
/etc/security ConfFiles
# PAM libraries
/lib(64)?/security Binlib

Keep the database offline and read-only

A second important aspect is that the result database should be stored off-line when not needed, and use it in read-only modus when the database is needed. This gives some protection against a malicious user, that might already have compromised the machine, to also modify the results database. For instance, provide the result database on a read-only NFS mount (for servers) or read-only medium (when physical access to the machine is possible) such as CD/DVD or read-only USB sticks.

After storing the database on such location, update the aide.conf file to have database= point to this new location.

Do offline scanning

If applicable, try using offline scanning methods for the system. In case of virtual platforms, it might be possible to take a snapshot of the system, mount this snapshot (read-only) and then run the aide scan on the mounted file system.

root #losetup /dev/loop0 /srv/virt/gentoo.img
root #vgscan
root #vgchange -ay
root #mount -o ro /dev/volgrpX/volumeY /mnt/image
root #chroot /mnt/image
root #aide --check --config=/path/to/aide.conf
root #exit
root #umount /mnt/image
root #vgchange -an /dev/volgrpX
root #losetup -d /dev/loop0

The above approach uses a chroot. This is only needed when the initial file system has been scanned from the live system and the administrator wants to perform an offline validation. If the initial scan was done offline, then the aide.conf file will point to the mount point already and the database will use these paths immediately, so then there is no need for chrooting.

See also