Those options mapped to the deprecated V1 format,
and didn't allow using the newer format.
If any of the older options are used,
a warning is displayed,
and the generated registries.conf file follows the old V1 format.
For more information about the new format,
see https://github.com/containers/image/blob/main/docs/containers-registries.conf.5.md
When softDaemonRestart is enabled, containers and virtual machines still get restarted when incus updates
Disabling stopIfChanged on both incus-startup and incus fixes this
In a13b423085 ("nixos-containers: allow hard-coding container veth MAC address &
add IPv6 SLAAC test case (#462775)") I had a typo in the `localMacAddress`
setting used for the dummy container config to generate the systemd service
template. I don't think this actually causes any breakage, but it's good to fix
for consistency's sake nonetheless.
The register-nix-paths service caused the nix database to be owned by
root, breaking unprivileged daemon support.
It is unclear why this worked previously.
In 9e78baf, Aarch64 machines were added the `-device virtio-gpu-pci`
argument, to mirror `-vga std` option that was added unconditionally to
virtualised x86_64 machines. This also enabled screenshots on aarch64
machines to be taken in tests. Since then, in a7ca287, the `-vga std`
option was removed, because that is the default value of this option
embedded in QEMU since version 2.2.
Removing it from the hard-coded list of qemu options makes it possible
for the user to choose their own `-vga` value. Similarly, for Aarch
virtualised machines, choosing unconditionally the `-device virtio-gpu-pci`
option instead of e.g `-device virtio-gpu-gl-pci`, can be bothersome too - see:
https://discourse.nixos.org/t/test-an-aarch64-linux-vm-wayland-compositor-from-an-x86-64-linux-host/77416/2
This commit is similar to 44c6c2ef16 which
was reverted in #527048.
The fix in #523016 fixes an issue with the default gateway, however it
does so by unconditionally defining `networking.interfaces.eth0`. This
makes so if you had other methods of defining the addresses for eth0,
those now get blanked out.
Instead, move the logic around so we only define
networking.interfaces.eth0 if we really have to.
Fixes issue reported here https://github.com/NixOS/nixpkgs/pull/515773#issuecomment-4501563586
Containers with privateNetwork have an eth0 interface configured
imperatively by the container setup script, so the networking-interfaces.nix module
doesn't know about it. Specifying the default gateway then fails
silently, unless this setup is mirrored in networking.interfaces.eth0
inside the container.
When "non-user" modules are not placed in baseModules or
extraModules (arguments of eval-config.nix), noUserModules.evalModules
breaks if the options defined in those modules are consumed. In order
to restore noUserModules functionality for NixOS VM (and now nspawn)
tests, the modules implementing test behavior are moved to baseModules.
In 9e78baf5c7, Aarch machines were added
the `-device virtio-gpu-pci` argument, to mirror `-vga std` option that
was added unconditionally to virtualised x86_64 machines. Since then,
in a7ca287ecb, the `-vga std` option was
removed, because that is the default value of this option embedded in
QEMU since version 2.2.
Removing it from the hard-coded list of qemu options makes it possible
for the user to choose their own `-vga` value. Similarly, for Aarch
virtualised machines, choosing unconditionally the `-device
virtio-gpu-pci` option instead of e.g `-device virtio-gpu-gl-pci`,
can be bothersome too - see:
https://discourse.nixos.org/t/test-an-aarch64-linux-vm-wayland-compositor-from-an-x86-64-linux-host/77416/2
readFile on a derivation output forces a build during evaluation.
Symlink the upstream file via environment.etc instead, matching the
adjacent sysctl.d handling.
Minijackson