commit 0d789e992fad80ee5b8c05c44fa0de746bf85594 · pyrox.dev/nixpkgs

+44

nixos/doc/manual/administration/container-networking.section.md

···

       1
       1
       +
       # Container Networking {#sec-container-networking}

     

       2
       2
       +
       

     

       3
       3
       +
       When you create a container using `nixos-container create`, it gets it

     

       4
       4
       +
       own private IPv4 address in the range `10.233.0.0/16`. You can get the

     

       5
       5
       +
       container's IPv4 address as follows:

     

       6
       6
       +
       

     

       7
       7
       +
       ```ShellSession

     

       8
       8
       +
       # nixos-container show-ip foo

     

       9
       9
       +
       10.233.4.2

     

       10
       10
       +
       

     

       11
       11
       +
       $ ping -c1 10.233.4.2

     

       12
       12
       +
       64 bytes from 10.233.4.2: icmp_seq=1 ttl=64 time=0.106 ms

     

       13
       13
       +
       ```

     

       14
       14
       +
       

     

       15
       15
       +
       Networking is implemented using a pair of virtual Ethernet devices. The

     

       16
       16
       +
       network interface in the container is called `eth0`, while the matching

     

       17
       17
       +
       interface in the host is called `ve-container-name` (e.g., `ve-foo`).

     

       18
       18
       +
       The container has its own network namespace and the `CAP_NET_ADMIN`

     

       19
       19
       +
       capability, so it can perform arbitrary network configuration such as

     

       20
       20
       +
       setting up firewall rules, without affecting or having access to the

     

       21
       21
       +
       host's network.

     

       22
       22
       +
       

     

       23
       23
       +
       By default, containers cannot talk to the outside network. If you want

     

       24
       24
       +
       that, you should set up Network Address Translation (NAT) rules on the

     

       25
       25
       +
       host to rewrite container traffic to use your external IP address. This

     

       26
       26
       +
       can be accomplished using the following configuration on the host:

     

       27
       27
       +
       

     

       28
       28
       +
       ```nix

     

       29
       29
       +
       networking.nat.enable = true;

     

       30
       30
       +
       networking.nat.internalInterfaces = ["ve-+"];

     

       31
       31
       +
       networking.nat.externalInterface = "eth0";

     

       32
       32
       +
       ```

     

       33
       33
       +
       

     

       34
       34
       +
       where `eth0` should be replaced with the desired external interface.

     

       35
       35
       +
       Note that `ve-+` is a wildcard that matches all container interfaces.

     

       36
       36
       +
       

     

       37
       37
       +
       If you are using Network Manager, you need to explicitly prevent it from

     

       38
       38
       +
       managing container interfaces:

     

       39
       39
       +
       

     

       40
       40
       +
       ```nix

     

       41
       41
       +
       networking.networkmanager.unmanaged = [ "interface-name:ve-*" ];

     

       42
       42
       +
       ```

     

       43
       43
       +
       

     

       44
       44
       +
       You may need to restart your system for the changes to take effect.

-59

nixos/doc/manual/administration/container-networking.xml

···

       1
       1
       -
       <section  xmlns="http://docbook.org/ns/docbook"

     

       2
       2
       -
                 xmlns:xlink="http://www.w3.org/1999/xlink"

     

       3
       3
       -
                 xmlns:xi="http://www.w3.org/2001/XInclude"

     

       4
       4
       -
                 version="5.0"

     

       5
       5
       -
                 xml:id="sec-container-networking">

     

       6
       6
       -
        <title>Container Networking</title>

     

       7
       7
       -
       

     

       8
       8
       -
        <para>

     

       9
       9
       -
         When you create a container using <literal>nixos-container create</literal>,

     

       10
       10
       -
         it gets it own private IPv4 address in the range

     

       11
       11
       -
         <literal>10.233.0.0/16</literal>. You can get the container’s IPv4 address

     

       12
       12
       -
         as follows:

     

       13
       13
       -
       <screen>

     

       14
       14
       -
       <prompt># </prompt>nixos-container show-ip foo

     

       15
       15
       -
       10.233.4.2

     

       16
       16
       -
       

     

       17
       17
       -
       <prompt>$ </prompt>ping -c1 10.233.4.2

     

       18
       18
       -
       64 bytes from 10.233.4.2: icmp_seq=1 ttl=64 time=0.106 ms

     

       19
       19
       -
       </screen>

     

       20
       20
       -
        </para>

     

       21
       21
       -
       

     

       22
       22
       -
        <para>

     

       23
       23
       -
         Networking is implemented using a pair of virtual Ethernet devices. The

     

       24
       24
       -
         network interface in the container is called <literal>eth0</literal>, while

     

       25
       25
       -
         the matching interface in the host is called

     

       26
       26
       -
         <literal>ve-<replaceable>container-name</replaceable></literal> (e.g.,

     

       27
       27
       -
         <literal>ve-foo</literal>). The container has its own network namespace and

     

       28
       28
       -
         the <literal>CAP_NET_ADMIN</literal> capability, so it can perform arbitrary

     

       29
       29
       -
         network configuration such as setting up firewall rules, without affecting or

     

       30
       30
       -
         having access to the host’s network.

     

       31
       31
       -
        </para>

     

       32
       32
       -
       

     

       33
       33
       -
        <para>

     

       34
       34
       -
         By default, containers cannot talk to the outside network. If you want that,

     

       35
       35
       -
         you should set up Network Address Translation (NAT) rules on the host to

     

       36
       36
       -
         rewrite container traffic to use your external IP address. This can be

     

       37
       37
       -
         accomplished using the following configuration on the host:

     

       38
       38
       -
       <programlisting>

     

       39
       39
       -
       <xref linkend="opt-networking.nat.enable"/> = true;

     

       40
       40
       -
       <xref linkend="opt-networking.nat.internalInterfaces"/> = ["ve-+"];

     

       41
       41
       -
       <xref linkend="opt-networking.nat.externalInterface"/> = "eth0";

     

       42
       42
       -
       </programlisting>

     

       43
       43
       -
         where <literal>eth0</literal> should be replaced with the desired external

     

       44
       44
       -
         interface. Note that <literal>ve-+</literal> is a wildcard that matches all

     

       45
       45
       -
         container interfaces.

     

       46
       46
       -
        </para>

     

       47
       47
       -
       

     

       48
       48
       -
        <para>

     

       49
       49
       -
         If you are using Network Manager, you need to explicitly prevent it from

     

       50
       50
       -
         managing container interfaces:

     

       51
       51
       -
       <programlisting>

     

       52
       52
       -
       networking.networkmanager.unmanaged = [ "interface-name:ve-*" ];

     

       53
       53
       -
       </programlisting>

     

       54
       54
       -
        </para>

     

       55
       55
       -
       

     

       56
       56
       -
        <para>

     

       57
       57
       -
         You may need to restart your system for the changes to take effect.

     

       58
       58
       -
        </para>

     

       59
       59
       -
       </section>

+3 -3

nixos/doc/manual/administration/containers.xml

···

       28
       28
        
         contrast, in the imperative approach, containers are configured and updated

     

       29
       29
        
         independently from the host system.

     

       30
       30
        
        </para>

     

       31
       31
       -
        <xi:include href="imperative-containers.xml" />

     

       32
       32
       -
        <xi:include href="declarative-containers.xml" />

     

       33
       33
       -
        <xi:include href="container-networking.xml" />

     

       31
       31
       +
        <xi:include href="../from_md/administration/imperative-containers.section.xml" />

     

       32
       32
       +
        <xi:include href="../from_md/administration/declarative-containers.section.xml" />

     

       33
       33
       +
        <xi:include href="../from_md/administration/container-networking.section.xml" />

     

       34
       34
        
       </chapter>

+48

nixos/doc/manual/administration/declarative-containers.section.md

···

       1
       1
       +
       # Declarative Container Specification {#sec-declarative-containers}

     

       2
       2
       +
       

     

       3
       3
       +
       You can also specify containers and their configuration in the host's

     

       4
       4
       +
       `configuration.nix`. For example, the following specifies that there

     

       5
       5
       +
       shall be a container named `database` running PostgreSQL:

     

       6
       6
       +
       

     

       7
       7
       +
       ```nix

     

       8
       8
       +
       containers.database =

     

       9
       9
       +
         { config =

     

       10
       10
       +
             { config, pkgs, ... }:

     

       11
       11
       +
             { services.postgresql.enable = true;

     

       12
       12
       +
             services.postgresql.package = pkgs.postgresql_9_6;

     

       13
       13
       +
             };

     

       14
       14
       +
         };

     

       15
       15
       +
       ```

     

       16
       16
       +
       

     

       17
       17
       +
       If you run `nixos-rebuild switch`, the container will be built. If the

     

       18
       18
       +
       container was already running, it will be updated in place, without

     

       19
       19
       +
       rebooting. The container can be configured to start automatically by

     

       20
       20
       +
       setting `containers.database.autoStart = true` in its configuration.

     

       21
       21
       +
       

     

       22
       22
       +
       By default, declarative containers share the network namespace of the

     

       23
       23
       +
       host, meaning that they can listen on (privileged) ports. However, they

     

       24
       24
       +
       cannot change the network configuration. You can give a container its

     

       25
       25
       +
       own network as follows:

     

       26
       26
       +
       

     

       27
       27
       +
       ```nix

     

       28
       28
       +
       containers.database = {

     

       29
       29
       +
         privateNetwork = true;

     

       30
       30
       +
         hostAddress = "192.168.100.10";

     

       31
       31
       +
         localAddress = "192.168.100.11";

     

       32
       32
       +
       };

     

       33
       33
       +
       ```

     

       34
       34
       +
       

     

       35
       35
       +
       This gives the container a private virtual Ethernet interface with IP

     

       36
       36
       +
       address `192.168.100.11`, which is hooked up to a virtual Ethernet

     

       37
       37
       +
       interface on the host with IP address `192.168.100.10`. (See the next

     

       38
       38
       +
       section for details on container networking.)

     

       39
       39
       +
       

     

       40
       40
       +
       To disable the container, just remove it from `configuration.nix` and

     

       41
       41
       +
       run `nixos-rebuild

     

       42
       42
       +
         switch`. Note that this will not delete the root directory of the

     

       43
       43
       +
       container in `/var/lib/containers`. Containers can be destroyed using

     

       44
       44
       +
       the imperative method: `nixos-container destroy foo`.

     

       45
       45
       +
       

     

       46
       46
       +
       Declarative containers can be started and stopped using the

     

       47
       47
       +
       corresponding systemd service, e.g.

     

       48
       48
       +
       `systemctl start container@database`.

-60

nixos/doc/manual/administration/declarative-containers.xml

···

       1
       1
       -
       <section  xmlns="http://docbook.org/ns/docbook"

     

       2
       2
       -
                 xmlns:xlink="http://www.w3.org/1999/xlink"

     

       3
       3
       -
                 xmlns:xi="http://www.w3.org/2001/XInclude"

     

       4
       4
       -
                 version="5.0"

     

       5
       5
       -
                 xml:id="sec-declarative-containers">

     

       6
       6
       -
        <title>Declarative Container Specification</title>

     

       7
       7
       -
       

     

       8
       8
       -
        <para>

     

       9
       9
       -
         You can also specify containers and their configuration in the host’s

     

       10
       10
       -
         <filename>configuration.nix</filename>. For example, the following specifies

     

       11
       11
       -
         that there shall be a container named <literal>database</literal> running

     

       12
       12
       -
         PostgreSQL:

     

       13
       13
       -
       <programlisting>

     

       14
       14
       -
       containers.database =

     

       15
       15
       -
         { config =

     

       16
       16
       -
             { config, pkgs, ... }:

     

       17
       17
       -
             { <xref linkend="opt-services.postgresql.enable"/> = true;

     

       18
       18
       -
             <xref linkend="opt-services.postgresql.package"/> = pkgs.postgresql_9_6;

     

       19
       19
       -
             };

     

       20
       20
       -
         };

     

       21
       21
       -
       </programlisting>

     

       22
       22
       -
         If you run <literal>nixos-rebuild switch</literal>, the container will be

     

       23
       23
       -
         built. If the container was already running, it will be updated in place,

     

       24
       24
       -
         without rebooting. The container can be configured to start automatically by

     

       25
       25
       -
         setting <literal>containers.database.autoStart = true</literal> in its

     

       26
       26
       -
         configuration.

     

       27
       27
       -
        </para>

     

       28
       28
       -
       

     

       29
       29
       -
        <para>

     

       30
       30
       -
         By default, declarative containers share the network namespace of the host,

     

       31
       31
       -
         meaning that they can listen on (privileged) ports. However, they cannot

     

       32
       32
       -
         change the network configuration. You can give a container its own network as

     

       33
       33
       -
         follows:

     

       34
       34
       -
       <programlisting>

     

       35
       35
       -
       containers.database = {

     

       36
       36
       -
         <link linkend="opt-containers._name_.privateNetwork">privateNetwork</link> = true;

     

       37
       37
       -
         <link linkend="opt-containers._name_.hostAddress">hostAddress</link> = "192.168.100.10";

     

       38
       38
       -
         <link linkend="opt-containers._name_.localAddress">localAddress</link> = "192.168.100.11";

     

       39
       39
       -
       };

     

       40
       40
       -
       </programlisting>

     

       41
       41
       -
         This gives the container a private virtual Ethernet interface with IP address

     

       42
       42
       -
         <literal>192.168.100.11</literal>, which is hooked up to a virtual Ethernet

     

       43
       43
       -
         interface on the host with IP address <literal>192.168.100.10</literal>. (See

     

       44
       44
       -
         the next section for details on container networking.)

     

       45
       45
       -
        </para>

     

       46
       46
       -
       

     

       47
       47
       -
        <para>

     

       48
       48
       -
         To disable the container, just remove it from

     

       49
       49
       -
         <filename>configuration.nix</filename> and run <literal>nixos-rebuild

     

       50
       50
       -
         switch</literal>. Note that this will not delete the root directory of the

     

       51
       51
       -
         container in <literal>/var/lib/containers</literal>. Containers can be

     

       52
       52
       -
         destroyed using the imperative method: <literal>nixos-container destroy

     

       53
       53
       -
         foo</literal>.

     

       54
       54
       -
        </para>

     

       55
       55
       -
       

     

       56
       56
       -
        <para>

     

       57
       57
       -
         Declarative containers can be started and stopped using the corresponding

     

       58
       58
       -
         systemd service, e.g. <literal>systemctl start container@database</literal>.

     

       59
       59
       -
        </para>

     

       60
       60
       -
       </section>

+115

nixos/doc/manual/administration/imperative-containers.section.md

···

       1
       1
       +
       # Imperative Container Management {#sec-imperative-containers}

     

       2
       2
       +
       

     

       3
       3
       +
       We'll cover imperative container management using `nixos-container`

     

       4
       4
       +
       first. Be aware that container management is currently only possible as

     

       5
       5
       +
       `root`.

     

       6
       6
       +
       

     

       7
       7
       +
       You create a container with identifier `foo` as follows:

     

       8
       8
       +
       

     

       9
       9
       +
       ```ShellSession

     

       10
       10
       +
       # nixos-container create foo

     

       11
       11
       +
       ```

     

       12
       12
       +
       

     

       13
       13
       +
       This creates the container's root directory in `/var/lib/containers/foo`

     

       14
       14
       +
       and a small configuration file in `/etc/containers/foo.conf`. It also

     

       15
       15
       +
       builds the container's initial system configuration and stores it in

     

       16
       16
       +
       `/nix/var/nix/profiles/per-container/foo/system`. You can modify the

     

       17
       17
       +
       initial configuration of the container on the command line. For

     

       18
       18
       +
       instance, to create a container that has `sshd` running, with the given

     

       19
       19
       +
       public key for `root`:

     

       20
       20
       +
       

     

       21
       21
       +
       ```ShellSession

     

       22
       22
       +
       # nixos-container create foo --config '

     

       23
       23
       +
         services.openssh.enable = true;

     

       24
       24
       +
         users.users.root.openssh.authorizedKeys.keys = ["ssh-dss AAAAB3N…"];

     

       25
       25
       +
       '

     

       26
       26
       +
       ```

     

       27
       27
       +
       

     

       28
       28
       +
       By default the next free address in the `10.233.0.0/16` subnet will be

     

       29
       29
       +
       chosen as container IP. This behavior can be altered by setting

     

       30
       30
       +
       `--host-address` and `--local-address`:

     

       31
       31
       +
       

     

       32
       32
       +
       ```ShellSession

     

       33
       33
       +
       # nixos-container create test --config-file test-container.nix \

     

       34
       34
       +
           --local-address 10.235.1.2 --host-address 10.235.1.1

     

       35
       35
       +
       ```

     

       36
       36
       +
       

     

       37
       37
       +
       Creating a container does not start it. To start the container, run:

     

       38
       38
       +
       

     

       39
       39
       +
       ```ShellSession

     

       40
       40
       +
       # nixos-container start foo

     

       41
       41
       +
       ```

     

       42
       42
       +
       

     

       43
       43
       +
       This command will return as soon as the container has booted and has

     

       44
       44
       +
       reached `multi-user.target`. On the host, the container runs within a

     

       45
       45
       +
       systemd unit called `container@container-name.service`. Thus, if

     

       46
       46
       +
       something went wrong, you can get status info using `systemctl`:

     

       47
       47
       +
       

     

       48
       48
       +
       ```ShellSession

     

       49
       49
       +
       # systemctl status container@foo

     

       50
       50
       +
       ```

     

       51
       51
       +
       

     

       52
       52
       +
       If the container has started successfully, you can log in as root using

     

       53
       53
       +
       the `root-login` operation:

     

       54
       54
       +
       

     

       55
       55
       +
       ```ShellSession

     

       56
       56
       +
       # nixos-container root-login foo

     

       57
       57
       +
       [root@foo:~]#

     

       58
       58
       +
       ```

     

       59
       59
       +
       

     

       60
       60
       +
       Note that only root on the host can do this (since there is no

     

       61
       61
       +
       authentication). You can also get a regular login prompt using the

     

       62
       62
       +
       `login` operation, which is available to all users on the host:

     

       63
       63
       +
       

     

       64
       64
       +
       ```ShellSession

     

       65
       65
       +
       # nixos-container login foo

     

       66
       66
       +
       foo login: alice

     

       67
       67
       +
       Password: ***

     

       68
       68
       +
       ```

     

       69
       69
       +
       

     

       70
       70
       +
       With `nixos-container run`, you can execute arbitrary commands in the

     

       71
       71
       +
       container:

     

       72
       72
       +
       

     

       73
       73
       +
       ```ShellSession

     

       74
       74
       +
       # nixos-container run foo -- uname -a

     

       75
       75
       +
       Linux foo 3.4.82 #1-NixOS SMP Thu Mar 20 14:44:05 UTC 2014 x86_64 GNU/Linux

     

       76
       76
       +
       ```

     

       77
       77
       +
       

     

       78
       78
       +
       There are several ways to change the configuration of the container.

     

       79
       79
       +
       First, on the host, you can edit

     

       80
       80
       +
       `/var/lib/container/name/etc/nixos/configuration.nix`, and run

     

       81
       81
       +
       

     

       82
       82
       +
       ```ShellSession

     

       83
       83
       +
       # nixos-container update foo

     

       84
       84
       +
       ```

     

       85
       85
       +
       

     

       86
       86
       +
       This will build and activate the new configuration. You can also specify

     

       87
       87
       +
       a new configuration on the command line:

     

       88
       88
       +
       

     

       89
       89
       +
       ```ShellSession

     

       90
       90
       +
       # nixos-container update foo --config '

     

       91
       91
       +
         services.httpd.enable = true;

     

       92
       92
       +
         services.httpd.adminAddr = "foo@example.org";

     

       93
       93
       +
         networking.firewall.allowedTCPPorts = [ 80 ];

     

       94
       94
       +
       '

     

       95
       95
       +
       

     

       96
       96
       +
       # curl http://$(nixos-container show-ip foo)/

     

       97
       97
       +
       <!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 3.2 Final//EN">…

     

       98
       98
       +
       ```

     

       99
       99
       +
       

     

       100
       100
       +
       However, note that this will overwrite the container's

     

       101
       101
       +
       `/etc/nixos/configuration.nix`.

     

       102
       102
       +
       

     

       103
       103
       +
       Alternatively, you can change the configuration from within the

     

       104
       104
       +
       container itself by running `nixos-rebuild switch` inside the container.

     

       105
       105
       +
       Note that the container by default does not have a copy of the NixOS

     

       106
       106
       +
       channel, so you should run `nix-channel --update` first.

     

       107
       107
       +
       

     

       108
       108
       +
       Containers can be stopped and started using `nixos-container

     

       109
       109
       +
         stop` and `nixos-container start`, respectively, or by using

     

       110
       110
       +
       `systemctl` on the container's service unit. To destroy a container,

     

       111
       111
       +
       including its file system, do

     

       112
       112
       +
       

     

       113
       113
       +
       ```ShellSession

     

       114
       114
       +
       # nixos-container destroy foo

     

       115
       115
       +
       ```

-123

nixos/doc/manual/administration/imperative-containers.xml

···

       1
       1
       -
       <section  xmlns="http://docbook.org/ns/docbook"

     

       2
       2
       -
                 xmlns:xlink="http://www.w3.org/1999/xlink"

     

       3
       3
       -
                 xmlns:xi="http://www.w3.org/2001/XInclude"

     

       4
       4
       -
                 version="5.0"

     

       5
       5
       -
                 xml:id="sec-imperative-containers">

     

       6
       6
       -
        <title>Imperative Container Management</title>

     

       7
       7
       -
       

     

       8
       8
       -
        <para>

     

       9
       9
       -
         We’ll cover imperative container management using

     

       10
       10
       -
         <command>nixos-container</command> first. Be aware that container management

     

       11
       11
       -
         is currently only possible as <literal>root</literal>.

     

       12
       12
       -
        </para>

     

       13
       13
       -
       

     

       14
       14
       -
        <para>

     

       15
       15
       -
         You create a container with identifier <literal>foo</literal> as follows:

     

       16
       16
       -
       <screen>

     

       17
       17
       -
       <prompt># </prompt>nixos-container create <replaceable>foo</replaceable>

     

       18
       18
       -
       </screen>

     

       19
       19
       -
         This creates the container’s root directory in

     

       20
       20
       -
         <filename>/var/lib/containers/<replaceable>foo</replaceable></filename> and a small configuration file

     

       21
       21
       -
         in <filename>/etc/containers/<replaceable>foo</replaceable>.conf</filename>. It also builds the

     

       22
       22
       -
         container’s initial system configuration and stores it in

     

       23
       23
       -
         <filename>/nix/var/nix/profiles/per-container/<replaceable>foo</replaceable>/system</filename>. You can

     

       24
       24
       -
         modify the initial configuration of the container on the command line. For

     

       25
       25
       -
         instance, to create a container that has <command>sshd</command> running,

     

       26
       26
       -
         with the given public key for <literal>root</literal>:

     

       27
       27
       -
       <screen>

     

       28
       28
       -
       <prompt># </prompt>nixos-container create <replaceable>foo</replaceable> --config '

     

       29
       29
       -
         <xref linkend="opt-services.openssh.enable"/> = true;

     

       30
       30
       -
         <link linkend="opt-users.users._name_.openssh.authorizedKeys.keys">users.users.root.openssh.authorizedKeys.keys</link> = ["ssh-dss AAAAB3N…"];

     

       31
       31
       -
       '

     

       32
       32
       -
       </screen>

     

       33
       33
       -
         By default the next free address in the <literal>10.233.0.0/16</literal> subnet will be chosen

     

       34
       34
       -
         as container IP. This behavior can be altered by setting <literal>--host-address</literal> and

     

       35
       35
       -
         <literal>--local-address</literal>:

     

       36
       36
       -
       <screen>

     

       37
       37
       -
       <prompt># </prompt>nixos-container create test --config-file test-container.nix \

     

       38
       38
       -
           --local-address 10.235.1.2 --host-address 10.235.1.1

     

       39
       39
       -
       </screen>

     

       40
       40
       -
        </para>

     

       41
       41
       -
       

     

       42
       42
       -
        <para>

     

       43
       43
       -
         Creating a container does not start it. To start the container, run:

     

       44
       44
       -
       <screen>

     

       45
       45
       -
       <prompt># </prompt>nixos-container start <replaceable>foo</replaceable>

     

       46
       46
       -
       </screen>

     

       47
       47
       -
         This command will return as soon as the container has booted and has reached

     

       48
       48
       -
         <literal>multi-user.target</literal>. On the host, the container runs within

     

       49
       49
       -
         a systemd unit called

     

       50
       50
       -
         <literal>container@<replaceable>container-name</replaceable>.service</literal>.

     

       51
       51
       -
         Thus, if something went wrong, you can get status info using

     

       52
       52
       -
         <command>systemctl</command>:

     

       53
       53
       -
       <screen>

     

       54
       54
       -
       <prompt># </prompt>systemctl status container@<replaceable>foo</replaceable>

     

       55
       55
       -
       </screen>

     

       56
       56
       -
        </para>

     

       57
       57
       -
       

     

       58
       58
       -
        <para>

     

       59
       59
       -
         If the container has started successfully, you can log in as root using the

     

       60
       60
       -
         <command>root-login</command> operation:

     

       61
       61
       -
       <screen>

     

       62
       62
       -
       <prompt># </prompt>nixos-container root-login <replaceable>foo</replaceable>

     

       63
       63
       -
       <prompt>[root@foo:~]#</prompt>

     

       64
       64
       -
       </screen>

     

       65
       65
       -
         Note that only root on the host can do this (since there is no

     

       66
       66
       -
         authentication). You can also get a regular login prompt using the

     

       67
       67
       -
         <command>login</command> operation, which is available to all users on the

     

       68
       68
       -
         host:

     

       69
       69
       -
       <screen>

     

       70
       70
       -
       <prompt># </prompt>nixos-container login <replaceable>foo</replaceable>

     

       71
       71
       -
       foo login: alice

     

       72
       72
       -
       Password: ***

     

       73
       73
       -
       </screen>

     

       74
       74
       -
         With <command>nixos-container run</command>, you can execute arbitrary

     

       75
       75
       -
         commands in the container:

     

       76
       76
       -
       <screen>

     

       77
       77
       -
       <prompt># </prompt>nixos-container run <replaceable>foo</replaceable> -- uname -a

     

       78
       78
       -
       Linux foo 3.4.82 #1-NixOS SMP Thu Mar 20 14:44:05 UTC 2014 x86_64 GNU/Linux

     

       79
       79
       -
       </screen>

     

       80
       80
       -
        </para>

     

       81
       81
       -
       

     

       82
       82
       -
        <para>

     

       83
       83
       -
         There are several ways to change the configuration of the container. First,

     

       84
       84
       -
         on the host, you can edit

     

       85
       85
       -
         <literal>/var/lib/container/<replaceable>name</replaceable>/etc/nixos/configuration.nix</literal>,

     

       86
       86
       -
         and run

     

       87
       87
       -
       <screen>

     

       88
       88
       -
       <prompt># </prompt>nixos-container update <replaceable>foo</replaceable>

     

       89
       89
       -
       </screen>

     

       90
       90
       -
         This will build and activate the new configuration. You can also specify a

     

       91
       91
       -
         new configuration on the command line:

     

       92
       92
       -
       <screen>

     

       93
       93
       -
       <prompt># </prompt>nixos-container update <replaceable>foo</replaceable> --config '

     

       94
       94
       -
         <xref linkend="opt-services.httpd.enable"/> = true;

     

       95
       95
       -
         <xref linkend="opt-services.httpd.adminAddr"/> = "foo@example.org";

     

       96
       96
       -
         <xref linkend="opt-networking.firewall.allowedTCPPorts"/> = [ 80 ];

     

       97
       97
       -
       '

     

       98
       98
       -
       

     

       99
       99
       -
       <prompt># </prompt>curl http://$(nixos-container show-ip <replaceable>foo</replaceable>)/

     

       100
       100
       -
       &lt;!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 3.2 Final//EN">…

     

       101
       101
       -
       </screen>

     

       102
       102
       -
         However, note that this will overwrite the container’s

     

       103
       103
       -
         <filename>/etc/nixos/configuration.nix</filename>.

     

       104
       104
       -
        </para>

     

       105
       105
       -
       

     

       106
       106
       -
        <para>

     

       107
       107
       -
         Alternatively, you can change the configuration from within the container

     

       108
       108
       -
         itself by running <command>nixos-rebuild switch</command> inside the

     

       109
       109
       -
         container. Note that the container by default does not have a copy of the

     

       110
       110
       -
         NixOS channel, so you should run <command>nix-channel --update</command>

     

       111
       111
       -
         first.

     

       112
       112
       -
        </para>

     

       113
       113
       -
       

     

       114
       114
       -
        <para>

     

       115
       115
       -
         Containers can be stopped and started using <literal>nixos-container

     

       116
       116
       -
         stop</literal> and <literal>nixos-container start</literal>, respectively, or

     

       117
       117
       -
         by using <command>systemctl</command> on the container’s service unit. To

     

       118
       118
       -
         destroy a container, including its file system, do

     

       119
       119
       -
       <screen>

     

       120
       120
       -
       <prompt># </prompt>nixos-container destroy <replaceable>foo</replaceable>

     

       121
       121
       -
       </screen>

     

       122
       122
       -
        </para>

     

       123
       123
       -
       </section>

+54

nixos/doc/manual/from_md/administration/container-networking.section.xml

···

       1
       1
       +
       <section xmlns="http://docbook.org/ns/docbook" xmlns:xlink="http://www.w3.org/1999/xlink" xml:id="sec-container-networking">

     

       2
       2
       +
         <title>Container Networking</title>

     

       3
       3
       +
         <para>

     

       4
       4
       +
           When you create a container using

     

       5
       5
       +
           <literal>nixos-container create</literal>, it gets it own private

     

       6
       6
       +
           IPv4 address in the range <literal>10.233.0.0/16</literal>. You can

     

       7
       7
       +
           get the container’s IPv4 address as follows:

     

       8
       8
       +
         </para>

     

       9
       9
       +
         <programlisting>

     

       10
       10
       +
       # nixos-container show-ip foo

     

       11
       11
       +
       10.233.4.2

     

       12
       12
       +
       

     

       13
       13
       +
       $ ping -c1 10.233.4.2

     

       14
       14
       +
       64 bytes from 10.233.4.2: icmp_seq=1 ttl=64 time=0.106 ms

     

       15
       15
       +
       </programlisting>

     

       16
       16
       +
         <para>

     

       17
       17
       +
           Networking is implemented using a pair of virtual Ethernet devices.

     

       18
       18
       +
           The network interface in the container is called

     

       19
       19
       +
           <literal>eth0</literal>, while the matching interface in the host is

     

       20
       20
       +
           called <literal>ve-container-name</literal> (e.g.,

     

       21
       21
       +
           <literal>ve-foo</literal>). The container has its own network

     

       22
       22
       +
           namespace and the <literal>CAP_NET_ADMIN</literal> capability, so it

     

       23
       23
       +
           can perform arbitrary network configuration such as setting up

     

       24
       24
       +
           firewall rules, without affecting or having access to the host’s

     

       25
       25
       +
           network.

     

       26
       26
       +
         </para>

     

       27
       27
       +
         <para>

     

       28
       28
       +
           By default, containers cannot talk to the outside network. If you

     

       29
       29
       +
           want that, you should set up Network Address Translation (NAT) rules

     

       30
       30
       +
           on the host to rewrite container traffic to use your external IP

     

       31
       31
       +
           address. This can be accomplished using the following configuration

     

       32
       32
       +
           on the host:

     

       33
       33
       +
         </para>

     

       34
       34
       +
         <programlisting language="bash">

     

       35
       35
       +
       networking.nat.enable = true;

     

       36
       36
       +
       networking.nat.internalInterfaces = [&quot;ve-+&quot;];

     

       37
       37
       +
       networking.nat.externalInterface = &quot;eth0&quot;;

     

       38
       38
       +
       </programlisting>

     

       39
       39
       +
         <para>

     

       40
       40
       +
           where <literal>eth0</literal> should be replaced with the desired

     

       41
       41
       +
           external interface. Note that <literal>ve-+</literal> is a wildcard

     

       42
       42
       +
           that matches all container interfaces.

     

       43
       43
       +
         </para>

     

       44
       44
       +
         <para>

     

       45
       45
       +
           If you are using Network Manager, you need to explicitly prevent it

     

       46
       46
       +
           from managing container interfaces:

     

       47
       47
       +
         </para>

     

       48
       48
       +
         <programlisting language="bash">

     

       49
       49
       +
       networking.networkmanager.unmanaged = [ &quot;interface-name:ve-*&quot; ];

     

       50
       50
       +
       </programlisting>

     

       51
       51
       +
         <para>

     

       52
       52
       +
           You may need to restart your system for the changes to take effect.

     

       53
       53
       +
         </para>

     

       54
       54
       +
       </section>

+60

nixos/doc/manual/from_md/administration/declarative-containers.section.xml

···

       1
       1
       +
       <section xmlns="http://docbook.org/ns/docbook" xmlns:xlink="http://www.w3.org/1999/xlink" xml:id="sec-declarative-containers">

     

       2
       2
       +
         <title>Declarative Container Specification</title>

     

       3
       3
       +
         <para>

     

       4
       4
       +
           You can also specify containers and their configuration in the

     

       5
       5
       +
           host’s <literal>configuration.nix</literal>. For example, the

     

       6
       6
       +
           following specifies that there shall be a container named

     

       7
       7
       +
           <literal>database</literal> running PostgreSQL:

     

       8
       8
       +
         </para>

     

       9
       9
       +
         <programlisting language="bash">

     

       10
       10
       +
       containers.database =

     

       11
       11
       +
         { config =

     

       12
       12
       +
             { config, pkgs, ... }:

     

       13
       13
       +
             { services.postgresql.enable = true;

     

       14
       14
       +
             services.postgresql.package = pkgs.postgresql_9_6;

     

       15
       15
       +
             };

     

       16
       16
       +
         };

     

       17
       17
       +
       </programlisting>

     

       18
       18
       +
         <para>

     

       19
       19
       +
           If you run <literal>nixos-rebuild switch</literal>, the container

     

       20
       20
       +
           will be built. If the container was already running, it will be

     

       21
       21
       +
           updated in place, without rebooting. The container can be configured

     

       22
       22
       +
           to start automatically by setting

     

       23
       23
       +
           <literal>containers.database.autoStart = true</literal> in its

     

       24
       24
       +
           configuration.

     

       25
       25
       +
         </para>

     

       26
       26
       +
         <para>

     

       27
       27
       +
           By default, declarative containers share the network namespace of

     

       28
       28
       +
           the host, meaning that they can listen on (privileged) ports.

     

       29
       29
       +
           However, they cannot change the network configuration. You can give

     

       30
       30
       +
           a container its own network as follows:

     

       31
       31
       +
         </para>

     

       32
       32
       +
         <programlisting language="bash">

     

       33
       33
       +
       containers.database = {

     

       34
       34
       +
         privateNetwork = true;

     

       35
       35
       +
         hostAddress = &quot;192.168.100.10&quot;;

     

       36
       36
       +
         localAddress = &quot;192.168.100.11&quot;;

     

       37
       37
       +
       };

     

       38
       38
       +
       </programlisting>

     

       39
       39
       +
         <para>

     

       40
       40
       +
           This gives the container a private virtual Ethernet interface with

     

       41
       41
       +
           IP address <literal>192.168.100.11</literal>, which is hooked up to

     

       42
       42
       +
           a virtual Ethernet interface on the host with IP address

     

       43
       43
       +
           <literal>192.168.100.10</literal>. (See the next section for details

     

       44
       44
       +
           on container networking.)

     

       45
       45
       +
         </para>

     

       46
       46
       +
         <para>

     

       47
       47
       +
           To disable the container, just remove it from

     

       48
       48
       +
           <literal>configuration.nix</literal> and run

     

       49
       49
       +
           <literal>nixos-rebuild switch</literal>. Note that this will not

     

       50
       50
       +
           delete the root directory of the container in

     

       51
       51
       +
           <literal>/var/lib/containers</literal>. Containers can be destroyed

     

       52
       52
       +
           using the imperative method:

     

       53
       53
       +
           <literal>nixos-container destroy foo</literal>.

     

       54
       54
       +
         </para>

     

       55
       55
       +
         <para>

     

       56
       56
       +
           Declarative containers can be started and stopped using the

     

       57
       57
       +
           corresponding systemd service, e.g.

     

       58
       58
       +
           <literal>systemctl start container@database</literal>.

     

       59
       59
       +
         </para>

     

       60
       60
       +
       </section>

+131

nixos/doc/manual/from_md/administration/imperative-containers.section.xml

···

       1
       1
       +
       <section xmlns="http://docbook.org/ns/docbook" xmlns:xlink="http://www.w3.org/1999/xlink" xml:id="sec-imperative-containers">

     

       2
       2
       +
         <title>Imperative Container Management</title>

     

       3
       3
       +
         <para>

     

       4
       4
       +
           We’ll cover imperative container management using

     

       5
       5
       +
           <literal>nixos-container</literal> first. Be aware that container

     

       6
       6
       +
           management is currently only possible as <literal>root</literal>.

     

       7
       7
       +
         </para>

     

       8
       8
       +
         <para>

     

       9
       9
       +
           You create a container with identifier <literal>foo</literal> as

     

       10
       10
       +
           follows:

     

       11
       11
       +
         </para>

     

       12
       12
       +
         <programlisting>

     

       13
       13
       +
       # nixos-container create foo

     

       14
       14
       +
       </programlisting>

     

       15
       15
       +
         <para>

     

       16
       16
       +
           This creates the container’s root directory in

     

       17
       17
       +
           <literal>/var/lib/containers/foo</literal> and a small configuration

     

       18
       18
       +
           file in <literal>/etc/containers/foo.conf</literal>. It also builds

     

       19
       19
       +
           the container’s initial system configuration and stores it in

     

       20
       20
       +
           <literal>/nix/var/nix/profiles/per-container/foo/system</literal>.

     

       21
       21
       +
           You can modify the initial configuration of the container on the

     

       22
       22
       +
           command line. For instance, to create a container that has

     

       23
       23
       +
           <literal>sshd</literal> running, with the given public key for

     

       24
       24
       +
           <literal>root</literal>:

     

       25
       25
       +
         </para>

     

       26
       26
       +
         <programlisting>

     

       27
       27
       +
       # nixos-container create foo --config '

     

       28
       28
       +
         services.openssh.enable = true;

     

       29
       29
       +
         users.users.root.openssh.authorizedKeys.keys = [&quot;ssh-dss AAAAB3N…&quot;];

     

       30
       30
       +
       '

     

       31
       31
       +
       </programlisting>

     

       32
       32
       +
         <para>

     

       33
       33
       +
           By default the next free address in the

     

       34
       34
       +
           <literal>10.233.0.0/16</literal> subnet will be chosen as container

     

       35
       35
       +
           IP. This behavior can be altered by setting

     

       36
       36
       +
           <literal>--host-address</literal> and

     

       37
       37
       +
           <literal>--local-address</literal>:

     

       38
       38
       +
         </para>

     

       39
       39
       +
         <programlisting>

     

       40
       40
       +
       # nixos-container create test --config-file test-container.nix \

     

       41
       41
       +
           --local-address 10.235.1.2 --host-address 10.235.1.1

     

       42
       42
       +
       </programlisting>

     

       43
       43
       +
         <para>

     

       44
       44
       +
           Creating a container does not start it. To start the container, run:

     

       45
       45
       +
         </para>

     

       46
       46
       +
         <programlisting>

     

       47
       47
       +
       # nixos-container start foo

     

       48
       48
       +
       </programlisting>

     

       49
       49
       +
         <para>

     

       50
       50
       +
           This command will return as soon as the container has booted and has

     

       51
       51
       +
           reached <literal>multi-user.target</literal>. On the host, the

     

       52
       52
       +
           container runs within a systemd unit called

     

       53
       53
       +
           <literal>container@container-name.service</literal>. Thus, if

     

       54
       54
       +
           something went wrong, you can get status info using

     

       55
       55
       +
           <literal>systemctl</literal>:

     

       56
       56
       +
         </para>

     

       57
       57
       +
         <programlisting>

     

       58
       58
       +
       # systemctl status container@foo

     

       59
       59
       +
       </programlisting>

     

       60
       60
       +
         <para>

     

       61
       61
       +
           If the container has started successfully, you can log in as root

     

       62
       62
       +
           using the <literal>root-login</literal> operation:

     

       63
       63
       +
         </para>

     

       64
       64
       +
         <programlisting>

     

       65
       65
       +
       # nixos-container root-login foo

     

       66
       66
       +
       [root@foo:~]#

     

       67
       67
       +
       </programlisting>

     

       68
       68
       +
         <para>

     

       69
       69
       +
           Note that only root on the host can do this (since there is no

     

       70
       70
       +
           authentication). You can also get a regular login prompt using the

     

       71
       71
       +
           <literal>login</literal> operation, which is available to all users

     

       72
       72
       +
           on the host:

     

       73
       73
       +
         </para>

     

       74
       74
       +
         <programlisting>

     

       75
       75
       +
       # nixos-container login foo

     

       76
       76
       +
       foo login: alice

     

       77
       77
       +
       Password: ***

     

       78
       78
       +
       </programlisting>

     

       79
       79
       +
         <para>

     

       80
       80
       +
           With <literal>nixos-container run</literal>, you can execute

     

       81
       81
       +
           arbitrary commands in the container:

     

       82
       82
       +
         </para>

     

       83
       83
       +
         <programlisting>

     

       84
       84
       +
       # nixos-container run foo -- uname -a

     

       85
       85
       +
       Linux foo 3.4.82 #1-NixOS SMP Thu Mar 20 14:44:05 UTC 2014 x86_64 GNU/Linux

     

       86
       86
       +
       </programlisting>

     

       87
       87
       +
         <para>

     

       88
       88
       +
           There are several ways to change the configuration of the container.

     

       89
       89
       +
           First, on the host, you can edit

     

       90
       90
       +
           <literal>/var/lib/container/name/etc/nixos/configuration.nix</literal>,

     

       91
       91
       +
           and run

     

       92
       92
       +
         </para>

     

       93
       93
       +
         <programlisting>

     

       94
       94
       +
       # nixos-container update foo

     

       95
       95
       +
       </programlisting>

     

       96
       96
       +
         <para>

     

       97
       97
       +
           This will build and activate the new configuration. You can also

     

       98
       98
       +
           specify a new configuration on the command line:

     

       99
       99
       +
         </para>

     

       100
       100
       +
         <programlisting>

     

       101
       101
       +
       # nixos-container update foo --config '

     

       102
       102
       +
         services.httpd.enable = true;

     

       103
       103
       +
         services.httpd.adminAddr = &quot;foo@example.org&quot;;

     

       104
       104
       +
         networking.firewall.allowedTCPPorts = [ 80 ];

     

       105
       105
       +
       '

     

       106
       106
       +
       

     

       107
       107
       +
       # curl http://$(nixos-container show-ip foo)/

     

       108
       108
       +
       &lt;!DOCTYPE HTML PUBLIC &quot;-//W3C//DTD HTML 3.2 Final//EN&quot;&gt;…

     

       109
       109
       +
       </programlisting>

     

       110
       110
       +
         <para>

     

       111
       111
       +
           However, note that this will overwrite the container’s

     

       112
       112
       +
           <literal>/etc/nixos/configuration.nix</literal>.

     

       113
       113
       +
         </para>

     

       114
       114
       +
         <para>

     

       115
       115
       +
           Alternatively, you can change the configuration from within the

     

       116
       116
       +
           container itself by running <literal>nixos-rebuild switch</literal>

     

       117
       117
       +
           inside the container. Note that the container by default does not

     

       118
       118
       +
           have a copy of the NixOS channel, so you should run

     

       119
       119
       +
           <literal>nix-channel --update</literal> first.

     

       120
       120
       +
         </para>

     

       121
       121
       +
         <para>

     

       122
       122
       +
           Containers can be stopped and started using

     

       123
       123
       +
           <literal>nixos-container stop</literal> and

     

       124
       124
       +
           <literal>nixos-container start</literal>, respectively, or by using

     

       125
       125
       +
           <literal>systemctl</literal> on the container’s service unit. To

     

       126
       126
       +
           destroy a container, including its file system, do

     

       127
       127
       +
         </para>

     

       128
       128
       +
         <programlisting>

     

       129
       129
       +
       # nixos-container destroy foo

     

       130
       130
       +
       </programlisting>

     

       131
       131
       +
       </section>