l2tp saves its secrets into /etc/ipsec.d but strongswan would not read
them. l2tp checks for /etc/ipsec.secrets includes /etc/ipsec.d and if
not tries to write into it.
Solution:
Have the strongswan module create /etc/ipsec.d and /etc/ipsec.secrets
when networkmanager_l2tp is installed.
Include /etc/ipsec.secrets in
/nix/store/hash-strongswan/etc/ipsec.secrets so that it can find l2tp
secrets.
Also when the ppp 'nopeerdns' option is used, the DNS resolver tries to
write into an alternate file /etc/ppp/resolv.conf. This fails when
/etc/ppp does not exist so the module creates it by default.
The use of Nix 2.0 significantly simplifies the installer, since we
can just pass a different store URI (--store /mnt) - it's no longer
needed to set up a chroot environment for the build, and to bootstrap
Nix into the chroot.
Also, commands that need to run in the installation (namely boot
loader installation and setting a root password) are now executed
using nixos-enter.
This also removes the need for nixos-prepare-root since any required
initialisation is done by Nix or by the activation script.
Regression introduced by 943592f698.
The lib attribute isn't in scope here, so we need to use pkgs.lib
instead for isFunction.
Signed-off-by: aszlig <aszlig@nix.build>
Cc: @shlevy
* The environment variables NIX_CONF_DIR, NIX_BUILD_HOOK and
NIX_REMOTE are no longer needed.
* A /bin/sh (from busybox) is provided by default in sandboxes.
* Various options were renamed.
Among other things, this will allow *2nix tools to output plain data
while still being composable with the traditional
callPackage/.override interfaces.
We had problems to get borg's own test suite running.
This test is intended to perform a quick smoke test to see whether we
have missed not any important dependency necessary to create backups
with borg.
tested with:
$ nix-build nixos/release.nix -A tests.borgbackup.x86_64-linux
This can be disabled with the `withKerberos` flag if desired.
Make the relevant assertions lazy,
so that if an overlay is used to set kerberos to null,
a later override can explicitly set `withKerberos` to false.
Don't build with GSSAPI by default;
the patchset is large and a bit hairy,
and it is reasonable to follow upstream who has not merged it
in not enabling it by default.
This can be disabled with the `withKerberos` flag if desired.
Make the relevant assertions lazy,
so that if an overlay is used to set kerberos to null,
a later override can explicitly set `withKerberos` to false.
Don't build with GSSAPI by default;
the patchset is large and a bit hairy,
and it is reasonable to follow upstream who has not merged it
in not enabling it by default.
`nixos-generate-config` detects the `cpuFreqGovernor` suited best for my
machine, e.g. `powerManagement.cpuFreqGovernor = lib.mkDefault "powersave";`.
However the `powerManagement` module sets a sensitive default for
`cpuFreqGovernor` using `mkDefault` to avoid breackage with older
setups. Since 140ac2f1 the `hardware-configuration.nix` sets the
gorvernor with `mkDefault` as well which causes evaluation errors if the
powermanagement module is enabled:
```
error: The unique option `powerManagement.cpuFreqGovernor' is defined multiple times, in `/home/ma27/Projects/nixos-config/hardware-configuration.nix' and `/nix/var/nix/profiles/per-user/root/channels/nixos/nixpkgs/nixos/modules/config/power-management.nix'.
```
Using `mkOptionDefault` rather than `mkDefault` in the powermanagement
module fixes this issue as it decreases the priority of the module and
prefers the value set in `hardware-configuration.nix`.
I have confirmed the change using the following VM declaration:
```
{
cpuFreq = { lib, ... }: {
powerManagement.cpuFreqGovernor = lib.mkDefault "powersave";
powerManagement.enable = true;
};
}
```
addPassthru became unused in #33057, but its signature was changed at the same
time. This commit restores the original signature and updates the warning and
the changelog.
`services.postfix.config` is now correctly merged with the default attrset
specified in the module. Some options that are lists in postfix also
have to be lists in nix to be merged correctly. Other default options are
now set with `mkDefault` so they can be overridden via the module system.
The existing callSubTests seems to already have special-cased code to
allow enabling subtests on a single specific system by looking at the
`system` attribute in the test arguments. Replace it with a new version
similar to the callTestOnTheseSystems because:
- It's consistent with the existing functions for creating
system-specific tests (though admittedly, the callSubTests special
case for `system` predates them)
- This approach allows limiting to multiple system types, the previous
one inherently allows only one system type.
- This also fixes the problem that if you pass in e.g.
supportedSystems = [ "aarch64-linux" ], you end up with a
tests.chromium job that silently runs on x86_64-linux.
- Finally, this causes renames of the jobs like:
tests.chromium -> tests.chromium.x86_64-linux to be consistent with
the rest of the tests.
- Add a new parameter `imageType` that can specify either "efi" or
"legacy" (the default which should see no change in behaviour by
this patch).
- EFI images get a GPT partition table (instead of msdos) with a
mandatory ESP partition (so we add an assert that `partitioned`
is true).
- Use the partx tool from util-linux to determine exact start + size
of the root partition. This is required because GPT stores a secondary
partition table at the end of the disk, so we can't just have
mkfs.ext4 create the filesystem until the end of the disk.
- (Unrelated to any EFI changes) Since we're depending on the
`-E offset=X` option to mkfs which is only supported by e2fsprogs,
disallow any attempts of creating partitioned disk images where
the root filesystem is not ext4.
New thin laptops don't have an ethernet port and so rely on wifi to get
access. With the minimal installer, setup wpa_supplicant can be hard if
it is the first time so here we provide an example.
Currently, even if you pass supportedSystems = [ "aarch64-linux" ] you
end up with e.g. `nixos.tests.docker` which actually silently runs on
x86_64-linux. Using the new callTestOnTheseSystems fixes that.
As a side-effect, this also causes a rename of
`nixos.tests.docker` -> `nixos.tests.docker.x86_64-linux`, which is IMHO
a good thing since it's makes them consistent with the rest of the
tests.
The default cache directory set by oh-my-zsh is $ohMyZsh/cache which
lives in the Nix store in our case. This causes issues with several
completion plugins provided by oh-my-zsh.
Currently, even if you pass `supportedSystems = [ "aarch64-linux" ]` you
end up with e.g. `nixos.iso_graphical.x86_64-linux` job. Using
forTheseSystems from release-lib avoids that.
This shouldn't affect the usual x86 trunk-combined jobset.
This was only applicable to very specific hardware, and the only person
with an apparent interest in maintaining it (me) no longer uses the
hardware in question.
When a domain has a lot of subdomains, it is quite easy to hit the rate limit:
https://letsencrypt.org/docs/rate-limits/
Instead you can define the certificate manually in `security.acme.certs` and list the subdomains in the `extraDomains` option.
sg and newgrp only changes the current user session and should be
available to users even if the "users.mutableUsers" option is set.
These are common, useful commands.
chfn does modify the /etc/passwd GECOS field which is also controlled
by the option "users.users.<name?>.description", so it's less
appropriate to make it available when "users.mutableUsers" is set.
However, because CHFN_RESTRICT in login.defs is never set in current
NixOS the chfn functionality is never available to users anyway and
may as well have its SUID disabled, as only root is able to use it.
This is recommended in the chfn man page in this case.
This makes memoization of Nixpkgs evaluation less effective, since
some Nixpkgs invocations may have 'config = {}' while others may have
'config = { xorg = {}; }'.
Instead set 'config = {}'.
This reverts commit 45c5a915980fbe1fa6f0ff80ab2d11b60b844d9e.
This breaks PredictableNetworkInterfaceNames on systems without networkd.
We should only include this file from systemd, when networkd is enabled.
in read-only way. If the cache directory is empty and you use the
very same service for system's DNS, kresd is unable to bootstrap root
trust anchors, as it would need a DNS lookup.
Also, if we don't rely on bootstrap, the extra lua deps of kresd could
be dropped by default, but let's not do that now, as the difference in
closure size is only ~4 MB, and there may be other use cases than
running the package as nixos service this way.
mate-control-center depends on mate-settings-daemon, but the later needs
gsettings schemas provided by the former. To fix this the gsettings schema
path from mate-control-center is added to XDG_DATA_DIRS at session
startup.
Udev changed its internal naming, so this rule file no longer applied correctly.
Therefore some properties such as network driver no longer matched in
systemd-networkd.
After updating we have more properties in systemd-networkd:
$ sudo networkctl status wlp3s0
...
Driver: iwlwifi
...
To prevent this in future, the file is no copied from systemd directly
The unnecessary dependency of sockets.target on kresd.service causes a
dependency cycle preventing kresd.service from starting at boot:
sockets.target -> kresd.service -> basic.target -> sockets.target
This allows to configure additional configuration files for Synapse. This way
secrets can be kept in a secure place on the file system without a need to go
through the Nix store.
"Ejecting" from the Finder ejects the entire device which is then not available for dd. diskutil unmountDisk does the right thing. Furthermore writing to diskN instead of rdiskN failed to complete even after waiting >10 minutes.
To make the configuration of `yabar` more pleasant and easier to
validate, a NixOS module will be quite helpful.
An example config could look like this:
```
{
programs.yabar = {
enable = true;
bars.top.indicators.exec = "YA_DATE";
};
}
```
The module adds a user-controlled systemd service which runs `yabar` after
starting up X.
before:
- /var/run/memcached is a bad default for a socket path, since its
parent directory must be writeable by memcached.
- Socket directory was not created by the module itself -> this was
left as a burden to the user?
- Having a static uid with a dynamic user name is not very useful.
after:
- Replace services.memcached.socket by a boolean flag. This simplifies
our code, since we do not have to check if the user specifies a
path with a parent directory that should be owned by memcached
(/run/memcached/memcached.sock -> /run/memcached).
- Remove fixed uid/gid allocation. The only file ever owned by the
daemon is the socket that will be recreated on every start.
Therefore user and group ids do not need to be static.
- only create the memcached user, if the user has not specified a
different one. The major use case for changing option is to allow
existing services (such as php-fpm) opening the local unix socket.
If we would unconditionally create a user that option would be
useless.
apps.plugin requires capabilities for full process monitoring. with
1.9.0, netdata allows multiple directories to search for plugins and the
setuid directory can be specified here.
the module is backwards compatible with older configs. a test is
included that verifies data gathering for the elevated privileges. one
additional attribute is added to make configuration more generic than
including configuration in string form.
It is quite complicated to test services using the test-driver when
declaring user services with `systemd.user.services` such as many
X11-based services like `xautolock.service`.
This change adds an optional `$user` parameter to each systemd-related
function in the test-driver and runs `systemctl --user` commands using
`su -l $user -c ...` and sets the `XDG_RUNTIME_DIR` variable
accordingly and a new function named `systemctl` which is able to run a
systemd command with or without a specified user.
The change can be confirmed with a simple VM declaration like this:
```
import ./nixos/tests/make-test.nix ({ pkgs, lib }:
with lib;
{
name = "systemd-user-test";
nodes.machine = {
imports = [ ./nixos/tests/common/user-account.nix ];
services.xserver.enable = true;
services.xserver.displayManager.auto.enable = true;
services.xserver.displayManager.auto.user = "bob";
services.xserver.xautolock.enable = true;
};
testScript = ''
$machine->start;
$machine->waitForX;
$machine->waitForUnit("xautolock.service", "bob");
$machine->stopJob("xautolock.service", "bob");
$machine->startJob("xautolock.service", "bob");
$machine->systemctl("list-jobs --no-pager", "bob");
$machine->systemctl("show 'xautolock.service' --no-pager", "bob");
'';
})
```
Instead of polluting the environment with environment variables which
are inherited by processes spawned from awesome, use the command line
argument "--search" to add things to the search path.
cc #33169
This enables znapzend users to specify its full configuration through
NixOS options, without ever needing to use the stateful `znapzendzetup`
command.
This works by running znapzendzetup with the specified config in
ExecPre, just before the znapzend daemon is started.
There is also the `pure` option which will clear all previous znapzend setups,
making it as stateless as can get, as only the setup declared in
configuration.nix will be persisted.
* bemenu: init at 2017-02-14
* velox: 2015-11-03 -> 2017-07-04
* orbment, velox: don't expose subprojects
the development of orbment and velox got stuck
their subprojects (bemenu, dmenu-wayland, st-wayland) don't work correctly outside of parent projects
so hide them to not confuse people
swc and wld libraries are unpopular and unlike wlc are not used by anything except velox
* pythonPackages.pydbus: init at 0.6.0
* way-cooler: 0.5.2 -> 0.6.2
* nixos/way-cooler: add module
* dconf module: use for wayland
non-invasive approach for #31293
see discussion at #32210
* sway: embed LD_LIBRARY_PATH for #32755
* way-cooler: switch from buildRustPackage to buildRustCrate #31150
This makes the commonHook option work also for (read-only) Nix store
paths. Currently it fails on the second activation, because the
destination is read-only.
Currently libvirt requires two qemu derivations: qemu and qemu_kvm which is just a truncated version of qemu (defined as qemu.override { hostCpuOnly = true; }).
This patch exposes an option virtualisation.libvirtd.qemuPackage which allows to choose which package to use:
* pkgs.qemu_kvm if all your guests have the same CPU as host, or
* pkgs.qemu which allows to emulate alien architectures (for example ARMV7L on X86_64), or
* a custom derivation
virtualisation.libvirtd.enableKVM option is vague and could be deprecate in favor of virtualisation.libvirtd.qemuPackage, anyway it does allow to enable/disable kvm.
(originally from f9415cb621)
feh is used to set background image for desktop managers that do not
support it directly, however there is no need to include it in PATH.
Fixes#17450.
Users were confused that the error message said config.networking.hostId, and indeed that did nothing to fix their problem.
Update the error message to specify the option they should actually set.
Now there are separate `xfce4.xfce4mixer_pulse` and `xfce4.xfcevolumed_pulse` attributes for PulseAudio versions of these packages, instead of relying on Nixpkgs option. Mind that xfce4-volumed and xfce4-volumed-pulse are actually two separate programs without much overlap.
Fakeroot seems to always give the owner write bit to any files touched
inside it (presumably to easily simulate the fact that root can still
modify such files). So do an explicit chmod to remove them.
This should finally solve #32242 after the EC2 images are regenerated
with this change.
https://hydra.nixos.org/build/66143116
Without this, when you've enabled networkmanager and start a
nixos-container the container will briefly have its specified IP
address but then networkmanager starts managing it causing the IP
address to be dropped.
* Add options:
- enable
- davUser (default: "davfs2")
- davGroup (default: "davfs2)
* Add davfs2 user or group if they are not specified in the
configuration
As described in detail here: https://github.com/NixOS/nixpkgs/issues/32533
bash will load completion scripts in $p/share/bash-completion/completions/ on
startup instead of letting bash-completion do it's lazy loading. Bash startup
will then slow down (very noticeable when bash-completion is installed in a
profile).
This commit leaves loading of scripts in the hands of bash-completion,
improving startup time for everyone using `enableCompletion`.
fixes#32533
In commit ec9dc73 restarting NetworkManager after resume from
suspend/hibernate was introduced.
When I initially switch to NixOS I started noticing a high delay between
wakeup and re-connecting to WiFi & wired networks. The delay increased
from a few seconds (on my previous distro, same software stack) to
almost half a minute with NixOS.
I (locally) applied the change in this commit a few weeks ago and tested
since then. The notebook/mobile device experience has improved a lot.
Reconnects are as before switching to NixOS.
Issue #24401 could be related to this. Since I am not using KDE/plasma5
I can only guess…
This allows one to specify "related packages" in NixOS that get rendered into
the configuration.nix(5) man page. The interface philosophy is pretty much
stolen from TeX bibliography.
This is required on the ThunderX CPUs on the Packet.net Type-2A
machines that have a GICv3. For some reason the default is to create a
GICv2 independent of the host hardware...
These packages will be placed into an environment using
`backendsToPackages`. This function explicitly maps backends to
`pkgs.nodePackages.${type}` unless it's a builtin. This ensures that only
valid backends that work on NixOS are used (if not, the build already
breaks at evaluation time).
The log will be redirected to `stdout` to be able to watch the entire
output using `journalctl`.
Configuration parameters for the backends need to be set using
`services.statsd.extraConfig` as each backend has its own options and
all of them shouldn't be validated and checked explicitly and manually.
This change adds a simple integration test exercising the fetchdocker
Nix code and hocker utilities for the simple `hello-world` docker
container. We exercise:
- Fetching the docker image configuration json
- Fetching the docker image layers
- Building a compositor script
- Loading the `hello-world` docker image into docker using the
compositor script and `docker load`
- Running that loaded container
Tunnel configuration has no member named "host" - i2pd does but it's called "address" in the options. As a result, no tunnel configuration is generated.
* Fix attribute check in inTunnels
* Fix integer to string coercion in inTunnels
* Add destinationPort option for outTunnels
Added the boolean option:
networking.networkmanager.enableStrongSwan
which enables the networkmanager_strongswan plugin and adds
strongswanNM to the dbus packages.
This was contributed by @wucke13, @eqyiel and @globin.
Fixes: #29873
There are security fixes in multiple packages /cc #32117,
so I'm merging a little earlier, with a few thousand jobs
still not finished on Hydra for x86_64-darwin and aarch64-linux.
- /var/run -> /run as the former is deprecated
- configure openal to use pulseaudio if pulseaudio is enabled
- configure libao to use pulseaudio if pulseaudio is enabled
For some reason, the GNOME 3.26 update broke the overrides. It turns
out the overrides now need to come before the overridden schemas in the
XDG_DATA_DIRS variable. This is not possible in general due to applications
prefixing the variable (e.g. in wrapGAppsHook).
To fix this, a new environment variable NIX_GSETTINGS_OVERRIDES_DIR
was introduced. It has greater priority than XDG_DATA_DIRS but lower
than GSETTINGS_SCHEMA_DIR. A separate variable was chosen in order not
to block the built-in one for users.
With libinput used for keyboard, base rules produce incorrect keyboard
layouts. We are removing the option as recommended in the XKB configuration
guide [1] to let X server choose the ruleset. It looks like it chooses
evdev rules which seem to work for some reason
[1]: https://www.x.org/releases/X11R7.6/doc/xorg-docs/input/XKB-Config.html#id2521360
The munin-node service used wrapProgram to inject environment variables.
This doesn't work because munin plugins depend on argv[0], which is
overwritten when the executable is a script with a shebang line (example
below).
This commit removes the wrappers and instead passes the required
environment variables to munin-node.
Eliminating the wrappers resulted in some broken plugins, e.g., meminfo
and hddtemp_smartctl. That was fixed with the per-plugin configuration.
Example:
The plugin if_eth0 is a symlink to /.../plugins/if_, which uses $0
to determine that it should monitor traffic on the eth0 interface.
if_ is a wrapped program, and runs `exec -a "$0" .if_-wrapped`
.if_-wrapped has a "#!/nix/.../bash" line, which results in bash
changing $0, and as a result the plugin thinks my interface
is called "-wrapped".
Changed extraUsers -> users and one case of extraGroups -> groups in nixos manual chapter 7.
According to chatter on IRC these are the proper names for these configuration options nowadays.
Modified based on feedback from Jörg Talheim.
Previously only the line numbers of a giant, internally generated XML file were
printed, without any kind of debuggability.
Now at least the mentioned lines are printed with a little bit of surrounding
context (to have something to grep for).
```
manual-combined.xml:4863: element para: Relax-NG validity error : Did not expect element para there
4859 <chapter xmlns="http://docbook.org/ns/docbook" xmlns:xlink="http://www.w3.org/1999/xlink" xmlns:xi="http://www.w3.org/2001/XInclude" version="5.0" xml:id="sec-writing-modules">
4860
4861 <title>Writing NixOS Modules</title>
4862
4863 <para>NixOS has a modular system for declarative configuration. This
4864 system combines multiple <emphasis>modules</emphasis> to produce the
4865 full system configuration. One of the modules that constitute the
```
The bash module currently sets the `/etc/inputrc` unconditionally,
which prevents easy user override. This commit lowers the priority of
the setting to "option default" level, which allows a user to override
the value using either
environment.etc."inputrc".text = …
or
environment.etc."inputrc".source = …
Fixes#28443
Fixed few invocations to `systemctl` to have an absolute path. Additionally add
LOCALE_ARCHIVE so that perl stops spewing warning messages.
