Docs: Add SELinux chapter

fixes #10553
This commit is contained in:
Dirk Goetz 2016-04-18 13:28:24 +02:00 committed by Michael Friedrich
parent 26d048d6c3
commit b2a1541394
8 changed files with 252 additions and 19 deletions

View File

@ -181,7 +181,7 @@ SNMP Traps can be received and filtered by using [SNMPTT](http://snmptt.sourcefo
and specific trap handlers passing the check results to Icinga 2.
Following the SNMPTT [Format](http://snmptt.sourceforge.net/docs/snmptt.shtml#SNMPTT.CONF-FORMAT)
documentation and the Icinga external command syntax found [here](23-appendix.md#external-commands-list-detail)
documentation and the Icinga external command syntax found [here](24-appendix.md#external-commands-list-detail)
we can create generic services that can accommodate any number of hosts for a given scenario.
### <a id="simple-traps"></a> Simple SNMP Traps

View File

@ -74,7 +74,7 @@ Example for PostgreSQL:
(1 Zeile)
A detailed list on the available table attributes can be found in the [DB IDO Schema documentation](23-appendix.md#schema-db-ido).
A detailed list on the available table attributes can be found in the [DB IDO Schema documentation](24-appendix.md#schema-db-ido).
## <a id="external-commands"></a> External Commands
@ -102,7 +102,7 @@ a forced service check:
Oct 17 15:01:25 icinga-server icinga2: Executing external command: [1382014885] SCHEDULE_FORCED_SVC_CHECK;localhost;ping4;1382014885
Oct 17 15:01:25 icinga-server icinga2: Rescheduling next check for service 'ping4'
A list of currently supported external commands can be found [here](23-appendix.md#external-commands-list-detail).
A list of currently supported external commands can be found [here](24-appendix.md#external-commands-list-detail).
Detailed information on the commands and their required parameters can be found
on the [Icinga 1.x documentation](http://docs.icinga.org/latest/en/extcommands2.html).
@ -432,7 +432,7 @@ re-implementation of the Livestatus protocol which is compatible with MK
Livestatus.
Details on the available tables and attributes with Icinga 2 can be found
in the [Livestatus Schema](23-appendix.md#schema-livestatus) section.
in the [Livestatus Schema](24-appendix.md#schema-livestatus) section.
You can enable Livestatus using icinga2 feature enable:
@ -508,7 +508,7 @@ Example using the tcp socket listening on port `6558`:
### <a id="livestatus-command-queries"></a> Livestatus COMMAND Queries
A list of available external commands and their parameters can be found [here](23-appendix.md#external-commands-list-detail)
A list of available external commands and their parameters can be found [here](24-appendix.md#external-commands-list-detail)
$ echo -e 'COMMAND <externalcommandstring>' | netcat 127.0.0.1 6558
@ -609,7 +609,7 @@ Default separators.
The `commands` table is populated with `CheckCommand`, `EventCommand` and `NotificationCommand` objects.
A detailed list on the available table attributes can be found in the [Livestatus Schema documentation](23-appendix.md#schema-livestatus).
A detailed list on the available table attributes can be found in the [Livestatus Schema documentation](24-appendix.md#schema-livestatus).
## <a id="status-data"></a> Status Data Files

232
doc/22-selinux.md Normal file
View File

@ -0,0 +1,232 @@
# <a id="selinux"></a> SELinux
## <a id="selinux-introduction"></a> Introduction
SELinux is a mandatory access control (MAC) system on Linux which adds a fine granular permission system for access to all resources on the system such as files, devices, networks and inter-process communication.
The most important questions are answered briefly in the [FAQ of the SELinux Project](http://selinuxproject.org/page/FAQ). For more details on SELinux and how to actually use and administrate it on your systems have a look at [Red Hat Enterprise Linux 7 - SELinux User's and Administrator's Guide](https://access.redhat.com/documentation/en-US/Red_Hat_Enterprise_Linux/7/html/SELinux_Users_and_Administrators_Guide/index.html). For an simplified (and funny) introduction download the [SELinux Coloring Book](https://github.com/mairin/selinux-coloring-book).
This documentation will use a similar format like the SELinux User's and Administrator's Guide.
### <a id="selinux-policy"></a> Policy
Icinga 2 is providing its own SELinux Policy. At the moment it is not upstreamed to the reference policy because it is under development. Target of the development is a policy package for Red Hat Enterprise Linux 7 and its derivates running the targeted policy which confines Icinga2 with all features and all checks executed. All other distributions will require some tweaks.
### <a id="selinux-policy-installation"></a> Installation
There are two ways to install the SELinux Policy for Icinga 2 on Enterprise Linux 7. Installing it from the provided package which is the preferred option and manual installation if you need some fixes not released yet or for development.
The policy package will run the daemon in a permissive domain so nothing will be denied also if the system runs in enforcing mode, so please make sure to run the system in this mode.
# sestatus
SELinux status: enabled
SELinuxfs mount: /sys/fs/selinux
SELinux root directory: /etc/selinux
Loaded policy name: targeted
Current mode: enforcing
Mode from config file: enforcing
Policy MLS status: enabled
Policy deny_unknown status: allowed
Max kernel policy version: 28
You can change the configured mode by editing `/etc/selinux/config` and the current mode by executing `setenforce 0`.
#### <a id="selinux-policy-installation-package"></a> Package installation
The packages are provided with release version 2.4 onwards.
Simply add the selinux subpackage to your installation.
# yum install icinga2-selinux
After that restart Icinga 2 and verify it running in its own domain `icinga2_t`.
# systemctl restart icinga2.service
# ps -eZ | grep icinga2
system_u:system_r:icinga2_t:s0 2825 ? 00:00:00 icinga2
#### <a id="selinux-policy-installation-manual"></a> Manual installation
This section describes the installation to support development and testing. It assumes that Icinga 2 is already installed from packages and running on the system.
As a prerequisite install the `git`, `selinux-policy-devel` and `audit` package. Enable and start the audit daemon afterwards.
# yum install git selinux-policy-devel audit
# systemctl enable auditd.service
# systemctl start auditd.service
After that clone the icinga2 git repository.
# git clone git://git.icinga.org/icinga2.git
To create and install the policy package run the installation script which also labels the resources. (The script assumes Icinga 2 was started once after system startup, the labeling of the port will only happen once and fail later on.)
# cd tools/selinux/
# ./icinga.sh
After that restart Icinga 2 and verify it running in its own domain `icinga2_t`.
# systemctl restart icinga2.service
# ps -eZ | grep icinga2
system_u:system_r:icinga2_t:s0 2825 ? 00:00:00 icinga2
### <a id="selinux-policy-general"></a> General
When the SELinux policy package for Icinga 2 is installed, the Icinga 2 daemon (icinga2) runs in its own domain `icinga2_t` and is separated from other confined services.
Files have to be labeled correctly for allowing icinga2 access to it. For example it writes to its own log files labeled `icinga2_log_t`. Also the API port is labeled `icinga_port_t` and icinga2 is allowed to manage it. Furthermore icinga2 can open high ports and unix sockets to connect to databases and features like graphite. It executes the nagios plugins and transitions to their context if those are labeled for example `nagios_services_plugin_exec_t` or `nagios_system_plugin_exec_t`.
Additional the Apache webserver is allowed to connect to the Command pipe of Icinga 2 to allow web interfaces sending commands to icinga2. This will perhaps change later on while investigating Icinga Web 2 for SELinux!
### <a id="selinux-policy-types"></a> Types
The command pipe is labeled `icinga2_command_t` and other services can request access to it by using the interface `icinga2_send_commands`.
The nagios plugins use their own contexts and icinga2 will transition to it. This means plugins have to be labeled correctly for their required permissions. The plugins installed from package should have set their permissions by the corresponding policy module and you can restore them using `restorecon -R -v /usr/lib64/nagios/plugins/`. To label your own plugins use `chcon -t type /path/to/plugin`, for the type have a look at table below.
Type | Domain | Use case | Provided by policy package
----------------------------------|------------------------------|------------------------------------------------------------------|---------------------------
nagios_admin_plugin_exec_t | nagios_admin_plugin_t | Plugins which require require read access on all file attributes | nagios
nagios_checkdisk_plugin_exec_t | nagios_checkdisk_plugin_t | Plugins which require read access to all filesystem attributes | nagios
nagios_mail_plugin_exec_t | nagios_mail_plugin_t | Plugins which access the local mail service | nagios
nagios_services_plugin_exec_t | nagios_services_plugin_t | Plugins monitoring network services | nagios
nagios_system_plugin_exec_t | nagios_system_plugin_t | Plugins checking local system state | nagios
nagios_unconfined_plugin_exec_t | nagios_unconfined_plugin_t | Plugins running without confinement | nagios
nagios_eventhandler_plugin_exec_t | nagios_eventhandler_plugin_t | Eventhandler (actually running unconfined) | nagios
nagios_openshift_plugin_exec_t | nagios_openshift_plugin_t | Plugins monitoring openshift | nagios
nagios_notification_plugin_exec_t | nagios_notification_plugin_t | Notification commands | icinga (will be moved later)
If one of those plugin domains causes problems you can set it to permissive by executing `semanage permissive -a domain`.
The policy provides a role `icinga2adm_r` for confining an user which enables an administrative user managing only Icinga 2 on the system. This user will also execute the plugins in their domain instead of the users one, so you can verify their execution with the same restrictions like they have when executed by icinga2.
### <a id="selinux-policy-booleans"></a> Booleans
SELinux is based on the least level of access required for a service to run. Using booleans you can grant more access in a defined way. The Icinga 2 policy package provides the following booleans.
**icinga2_can_connect_all**
Having this boolean enabled allows icinga2 to connect to all ports. This can be neccesary if you use features which connect to unconfined services.
### <a id="selinux-policy-examples"></a> Configuration Examples
#### <a id="selinux-policy-examples-plugin"></a> Confining a plugin
Download and install a plugin, for example check_mysql_health.
# wget http://labs.consol.de/download/shinken-nagios-plugins/check_mysql_health-2.1.9.2.tar.gz
# tar xvzf check_mysql_health-2.1.9.2.tar.gz
# cd check_mysql_health-2.1.9.2/
# ./configure --libexecdir /usr/lib64/nagios/plugins
# make
# make install
It is labeled `nagios_unconfined_plugins_exec_t` by default, so it runs without restrictions.
# ls -lZ /usr/lib64/nagios/plugins/check_mysql_health
-rwxr-xr-x. root root system_u:object_r:nagios_unconfined_plugin_exec_t:s0 /usr/lib64/nagios/plugins/check_mysql_health
In this case the plugin is monitoring a service, so it should be labeled `nagios_services_plugin_exec_t` to restrict its permissions.
# chcon -t nagios_services_plugin_exec_t /usr/lib64/nagios/plugins/check_mysql_health
# ls -lZ /usr/lib64/nagios/plugins/check_mysql_health
-rwxr-xr-x. root root system_u:object_r:nagios_services_plugin_exec_t:s0 /usr/lib64/nagios/plugins/check_mysql_health
The plugin still runs fine but if someone changes the script to do weird stuff it will fail to do so.
#### <a id="selinux-policy-examples-connectall"></a> Allow icinga to connect to all ports.
You are running graphite on a different port than `2003` and want `icinga2` to connect to it.
Change the port value for the graphite feature according to your graphite installation before enabling it.
# cat /etc/icinga2/features-enabled/graphite.conf
/**
* The GraphiteWriter type writes check result metrics and
* performance data to a graphite tcp socket.
*/
library "perfdata"
object GraphiteWriter "graphite" {
//host = "127.0.0.1"
//port = 2003
port = 2004
}
# icinga2 feature enable graphite
Before you restart the icinga2 service allow it to connect to all ports by enabling the boolean ´icinga2_can_connect_all` (now and permanent).
# setsebool icinga2_can_connect_all true
# setsebool -P icinga2_can_connect_all true
If you restart the daemon now it will successfully connect to graphite.
#### <a id="selinux-policy-examples-user"></a> Confining a user
If you want to have an administrative account capable of only managing icinga2 and not the complete system, you can restrict the privileges by confining
this user. This is completly optional!
Start by adding the Icinga 2 administrator role `icinga2adm_r` to the administrative SELinux user `staff_u`.
# semanage user -m -R "staff_r sysadm_r system_r unconfined_r icinga2adm_r" staff_u
Confine your user login and create a sudo rule.
# semanage login -a dirk -s staff_u
# echo "dirk ALL=(ALL) NOPASSWD: ALL" > /etc/sudoers.d/dirk
Login to the system using ssh and verify your id.
$ id -Z
staff_u:staff_r:staff_t:s0-s0:c0.c1023
Try to execute some commands as root using sudo.
$ sudo id -Z
staff_u:staff_r:staff_t:s0-s0:c0.c1023
$ sudo vi /etc/icinga2/icinga2.conf
"/etc/icinga2/icinga2.conf" [Permission Denied]
$ sudo cat /var/log/icinga2/icinga2.log
cat: /var/log/icinga2/icinga2.log: Keine Berechtigung
$ sudo systemctl reload icinga2.service
Failed to get D-Bus connection: No connection to service manager.
Those commands fail because you only switch to root but do not change your SELinux role. Try again but tell sudo also to switch the SELinux role and type.
$ sudo -r icinga2adm_r -t icinga2adm_t id -Z
staff_u:icinga2adm_r:icinga2adm_t:s0-s0:c0.c1023
$ sudo -r icinga2adm_r -t icinga2adm_t vi /etc/icinga2/icinga2.conf
"/etc/icinga2/icinga2.conf"
$ sudo -r icinga2adm_r -t icinga2adm_t cat /var/log/icinga2/icinga2.log
[2015-03-26 20:48:14 +0000] information/DynamicObject: Dumping program state to file '/var/lib/icinga2/icinga2.state'
$ sudo -r icinga2adm_r -t icinga2adm_t systemctl reload icinga2.service
Now the commands will work, but you have always to remember to add the arguments, so change the sudo rule to set it by default.
# echo "dirk ALL=(ALL) ROLE=icinga2adm_r TYPE=icinga2adm_t NOPASSWD: ALL" > /etc/sudoers.d/dirk
Now try the commands again without providing the role and type and they will work, but if you try to read apache logs or restart apache for example it will still fail.
$ sudo cat /var/log/httpd/error_log
/bin/cat: /var/log/httpd/error_log: Keine Berechtigung
$ sudo systemctl reload httpd.service
Failed to issue method call: Access denied
## <a id="selinux-bugreports"></a> Bugreports
If you experience any problems while running in enforcing mode try to reproduce it in permissive mode. If the problem persists it is not related to SELinux because in permissive mode SELinux will not deny anything.
For now Icinga 2 is running in a permissive domain and adds also some rules for other necessary services so no problems should occure at all. But you can help to enhance the policy by testing Icinga 2 running confined by SELinux.
When filing a bug report please add the following information additionally to the [normal ones](https://www.icinga.org/icinga/faq/):
* Output of `semodule -l | grep -e icinga2 -e nagios -e apache`
* Output of `ps -eZ | grep icinga2`
* Output of `semanage port -l | grep icinga2`
* Output of `audit2allow -li /var/log/audit/audit.log`
If access to a file is blocked and you can tell which one please provided the output of `ls -lZ /path/to/file` (and perhaps the directory above).
If asked for full audit.log add `-w /etc/shadow -p w` to `/etc/audit/rules.d/audit.rules`, restart the audit daemon, reproduce the problem and add `/var/log/audit/audit.log` to the bug report. With the added audit rule it will include the path of files access was denied to.
If asked to provide full audit log with dontaudit rules disabled executed `semodule -DB` before reproducing the problem. After that enable the rules again to prevent auditd spamming your logfile by executing `semodule -B`.

View File

@ -11,7 +11,7 @@ on your migration requirements.
For a long-term migration of your configuration you should consider re-creating
your configuration based on the proposed Icinga 2 configuration paradigm.
Please read the [next chapter](22-migrating-from-icinga-1x.md#differences-1x-2) to find out more about the differences
Please read the [next chapter](23-migrating-from-icinga-1x.md#differences-1x-2) to find out more about the differences
between 1.x and 2.
### <a id="manual-config-migration-hints"></a> Manual Config Migration Hints
@ -24,7 +24,7 @@ The examples are taken from Icinga 1.x test and production environments and conv
straight into a possible Icinga 2 format. If you found a different strategy, please
let us know!
If you require in-depth explanations, please check the [next chapter](22-migrating-from-icinga-1x.md#differences-1x-2).
If you require in-depth explanations, please check the [next chapter](23-migrating-from-icinga-1x.md#differences-1x-2).
#### <a id="manual-config-migration-hints-Intervals"></a> Manual Config Migration Hints for Intervals
@ -187,7 +187,7 @@ While you could manually migrate this like (please note the new generic command
#### <a id="manual-config-migration-hints-runtime-macros"></a> Manual Config Migration Hints for Runtime Macros
Runtime macros have been renamed. A detailed comparison table can be found [here](22-migrating-from-icinga-1x.md#differences-1x-2-runtime-macros).
Runtime macros have been renamed. A detailed comparison table can be found [here](23-migrating-from-icinga-1x.md#differences-1x-2-runtime-macros).
For example, accessing the service check output looks like the following in Icinga 1.x:
@ -260,7 +260,7 @@ while the service check command resolves its value to the service attribute attr
#### <a id="manual-config-migration-hints-contacts-users"></a> Manual Config Migration Hints for Contacts (Users)
Contacts in Icinga 1.x act as users in Icinga 2, but do not have any notification commands specified.
This migration part is explained in the [next chapter](22-migrating-from-icinga-1x.md#manual-config-migration-hints-notifications).
This migration part is explained in the [next chapter](23-migrating-from-icinga-1x.md#manual-config-migration-hints-notifications).
define contact{
contact_name testconfig-user
@ -270,8 +270,8 @@ This migration part is explained in the [next chapter](22-migrating-from-icinga-
email icinga@localhost
}
The `service_notification_options` can be [mapped](22-migrating-from-icinga-1x.md#manual-config-migration-hints-notification-filters)
into generic `state` and `type` filters if additional notification filtering is required. `alias` gets
The `service_notification_options` can be [mapped](23-migrating-from-icinga-1x.md#manual-config-migration-hints-notification-filters)
into generic `state` and `type` filters, if additional notification filtering is required. `alias` gets
renamed to `display_name`.
object User "testconfig-user" {
@ -322,7 +322,7 @@ Assign it to the host or service and set the newly generated notification comman
Convert the `notification_options` attribute from Icinga 1.x to Icinga 2 `states` and `types`. Details
[here](22-migrating-from-icinga-1x.md#manual-config-migration-hints-notification-filters). Add the notification period.
[here](23-migrating-from-icinga-1x.md#manual-config-migration-hints-notification-filters). Add the notification period.
states = [ OK, Warning, Critical ]
types = [ Recovery, Problem, Custom ]
@ -559,7 +559,7 @@ enabled.
assign where "hg_svcdep2" in host.groups
}
Host dependencies are explained in the [next chapter](22-migrating-from-icinga-1x.md#manual-config-migration-hints-host-parents).
Host dependencies are explained in the [next chapter](23-migrating-from-icinga-1x.md#manual-config-migration-hints-host-parents).
@ -930,7 +930,7 @@ In Icinga 1.x arguments are specified in the `check_command` attribute and
are separated from the command name using an exclamation mark (`!`).
Please check the migration hints for a detailed
[migration example](22-migrating-from-icinga-1x.md#manual-config-migration-hints-check-command-arguments).
[migration example](23-migrating-from-icinga-1x.md#manual-config-migration-hints-check-command-arguments).
> **Note**
>

View File

@ -1779,7 +1779,7 @@ Rephrased: If the parent service object changes into the `Warning` state, this
dependency will fail and render all child objects (hosts or services) unreachable.
You can determine the child's reachability by querying the `is_reachable` attribute
in for example [DB IDO](23-appendix.md#schema-db-ido-extensions).
in for example [DB IDO](24-appendix.md#schema-db-ido-extensions).
### <a id="dependencies-implicit-host-service"></a> Implicit Dependencies for Services on Host

View File

@ -678,5 +678,5 @@ safely reload the Icinga 2 daemon.
> which will validate the configuration in a separate process and not stop
> the other events like check execution, notifications, etc.
>
> Details can be found [here](22-migrating-from-icinga-1x.md#differences-1x-2-real-reload).
> Details can be found [here](23-migrating-from-icinga-1x.md#differences-1x-2-real-reload).

View File

@ -23,8 +23,9 @@ pages:
- [19-library-reference.md, Library Reference]
- [20-script-debugger.md, Script Debugger]
- [21-development.md, Development]
- [22-migrating-from-icinga-1x.md, Migrating from Icinga 1.x]
- [23-appendix.md, Appendix]
- [22-selinux.md, SELinux]
- [23-migrating-from-icinga-1x.md, Migrating from Icinga 1.x]
- [24-appendix.md, Appendix]
theme: readthedocs
markdown_extensions: [smarty]
extra_javascript: [scroll.js]