When using the declarative shared folder configuration for resilio sync
it is now possible to pass a path from which to read the secret should
be read at runtime. The path will not be added to the nix store.
The 'secret' parameter to specify the secret directly is still
supported. This option will still store the secret in the nix store.
This commit follows the pattern described in this issue, for upstream
programs that do not provide support for setting a password using a
file: https://github.com/NixOS/nixpkgs/issues/24288
The inline comments explain the reasoning behind this change. This
work was initiated due to failing tests explicitly for glusterfs, but
my hunch is that any nixosTest adjacent to rpcbind will start working
again.
Ref: https://github.com/NixOS/nixpkgs/issues/175339
This commit fixes broken non-declarative configs by
making the assertions more relaxed.
It also allows to remove the forced configuration merge by making
`settings` `null`able (now the default).
Both cases (trivial non-declarative config and `null`able config) are
verified with additional tests.
Fixes#198665
Relative paths are interpreted relative to the working directory, which
is currently unset and thus defaults to `/`. However we want to change
the working directory in a future release such that relative paths are
interpreted relative to `/var/lib/syncthing`.
in some circumstances, the setup service would fail with:
ERROR 1062 (23000) at line 5: Duplicate entry '1' for key 'PRIMARY'
so let's use an upsert instead of insert. This also simplifies the
script.
Besides, also fix that when the setup script changes, the corresponding
script is not restarted as it is usually not active, so we trigger a
restart of the main systemd service.
when a new peer is added, it does not modify any active units, because
the interface unit remains the same. therefore the new peer is not added
until next reboot or manual action.
Previously, dhcpcd and bitcoind starting up in parallel could lead to
the following error in bitcoind:
```
bitcoind: libevent: getaddrinfo: address family for nodename not supported
bitcoind: Binding RPC on address 127.0.0.1 port 8332 failed.
bitcoind: Unable to bind any endpoint for
```
After the initial failure, the bitcoind service would always restart successfully.
This race condition, where both applications were simultaneously
manipulating network resources, was only triggered under specific
hardware conditions.
Fix it by running bitcoind after dhcp has started (by running after
`network-online.target`).
This bug and the fix only affect the default NixOS scripted
networking backend.
If folders and devices are not configured explicitly, do not wipe the
changes done via the web GUI. Currently the list of devices or folders
will be reset unless overrideFolders/overrideDevices is disabled.
Make the dynamic-dns refresh systemd service (controlled via the
preexisting option dynamicEndpointRefreshSecond) robust to e.g. dns
failures that happen on intermittent network connections.
Background:
When dns resolution fails with a 'permanent' error ("Name or service not
known" instead of "Temporary failure in name resolution"), wireguard
won't retry despite WG_ENDPOINT_RESOLUTION_RETRIES=infinity.
-> This change should improve reliability/connectivity.
somewhat related thread: https://github.com/NixOS/nixpkgs/issues/63869
This will add `passthru.schema_version` to be used as default value for
the adguardhome module.
It will also update the `update.sh` to keep the `schema_version` in sync
with the version by inspecting the sourcecode.
This might break existing configs, if they use deprecated values that don't
appear in newer schema_versions and schema_version wasn't set explicitly.
Explicit declarations of schema_version always have higher priority.
This also removes the `host` and `config` settings in favour of using the
appropriate `settings`.
Fixes#173938
Co-authored-by: Sandro <sandro.jaeckel@gmail.com>
*Flags implies a list
slightly relevant:
> stdenv: start deprecating non-list configureFlags https://github.com/NixOS/nixpkgs/pull/173172
the makeInstalledTests function in `nixos/tests/installed-tests/default.nix` isn't available outside of nixpkgs so
it's not a breaking change
The original implementation had a few issues:
* The secret was briefly leaked since it is part of the cmdline for
`sed(1)` and on Linux `cmdline` is world-readable.
* If the secret would contain either a `,` or a `"` it would mess with
the `sed(1)` expression itself unless you apply messy escape hacks.
To circumvent all of that, I decided to use `replace-secret` which
allows you to replace a string inside a file (in this case
`#static-auth-secret#`) with the contents of a file, i.e.
`cfg.static-auth-secret-file` without any of these issues.
Some networks can only transfer packets with a lower than normal maximum
transfer unit size. In these cases, it is necessary to set a MTU that
works for the given upstream network.
Wireguard can tag its packets with a firewall mark. This can be used for
firewalls or policy routing. This is very useful in some setups where
all traffic should go through a wireguard interface. The wireguard
packets cannot go through the wireguard interface and must be routed
differently, which can be done via the Firewall Mark.
The nixos option `config.networking.wireguard.interface.<name>.fwMark`
is of type `types.str` and not `types.int` to allow for specifying the
mark as a hexadecimal value.
most of these are hidden because they're either part of a submodule that
doesn't have its type rendered (eg because the submodule type is used in
an either type) or because they are explicitly hidden. some of them are
merely hidden from nix-doc-munge by how their option is put together.
conversions were done using https://github.com/pennae/nix-doc-munge
using (probably) rev f34e145 running
nix-doc-munge nixos/**/*.nix
nix-doc-munge --import nixos/**/*.nix
the tool ensures that only changes that could affect the generated
manual *but don't* are committed, other changes require manual review
and are discarded.
there are sufficiently few variable list around, and they are
sufficiently simple, that it doesn't seem helpful to add another
markdown extension for them. rendering differences are small, except in
the tor module: admonitions inside other blocks cannot be made to work
well with mistune (and likely most other markdown processors), so those
had to be shuffled a bit. we also lose paragraph breaks in the list
items due to how we have to render from markdown to docbook, but once we
remove docbook from the pipeline those paragraph breaks will be restored.
mostly no rendering changes. some lists (like simplelist) don't have an
exact translation to markdown, so we use a comma-separated list of
literals instead.
this mostly means marking options that use markdown already
appropriately and making a few adjustments so they still render
correctly. notable for nftables we have to transform the md links
because the manpage would not render them correctly otherwise.
most of the screen tags used in option docs are actually listings of
some sort. nsd had a notable exception where its screen usage was pretty
much a raw markdown block that made most sense to convert into docbook lists.
Syncthing config XML uses `fsPath` setting for specifying the path to the versioning folder. This commit adds `services.syncthing.folders.<name>.versioning.fsPath` option to enable this functionality declaratively. Previously, `versioning.params.versionsPath` was used, which doesn't work.
using regular strings works well for docbook because docbook is not as
whitespace-sensitive as markdown. markdown would render all of these as
code blocks when given the chance.
a lot of markdown syntax has already snuck into option docs, many of it
predating the intent to migrate to markdown. we don't convert all of it
here, just that which is accompanied by docbook tags as well. the rest
can be converted by simply adding the mdDoc marker.
this renders the same in the manpage and a little more clearly in the
html manual. in the manpage there continues to be no distinction from
regular text, the html manual gets code-type markup (which was probably
the intention for most of these uses anyway).
now nix-doc-munge will not introduce whitespace changes when it replaces
manpage references with the MD equivalent.
no change to the manpage, changes to the HTML manual are whitespace only.
make (almost) all links appear on only a single line, with no
unnecessary whitespace, using double quotes for attributes. this lets us
automatically convert them to markdown easily.
the few remaining links are extremely long link in a gnome module, we'll
come back to those at a later date.
we can't embed syntactic annotations of this kind in markdown code
blocks without yet another extension. replaceable is rare enough to make
this not much worth it, so we'll go with «thing» instead. the module
system already uses this format for its placeholder names in attrsOf
paths.
markdown can't represent the difference without another extension and
both the html manual and the manpage render them the same, so keeping the
distinction is not very useful on its own. with the distinction removed
we can automatically convert many options that use <code> tags to markdown.
the manpage remains unchanged, html manual does not render
differently (but class names on code tags do change from "code" to "literal").
our xslt already replaces double line breaks with a paragraph close and
reopen. not using explicit para tags lets nix-doc-munge convert more
descriptions losslessly.
only whitespace changes to generated documents, except for two
strongswan options gaining paragraph two breaks they arguably should've
had anyway.
Enabling soju without providing a value for tlsCertificate currently
results in:
error: The option `services.soju.tlsCertificate' is used but not
defined.
Since tlsCertificate is intended to be optional, set default to null.
Additionally, add assertions to ensure that both tlsCertificate and
tlsCertificateKey are either set or unset.
the conversion procedure is simple:
- find all things that look like options, ie calls to either `mkOption`
or `lib.mkOption` that take an attrset. remember the attrset as the
option
- for all options, find a `description` attribute who's value is not a
call to `mdDoc` or `lib.mdDoc`
- textually convert the entire value of the attribute to MD with a few
simple regexes (the set from mdize-module.sh)
- if the change produced a change in the manual output, discard
- if the change kept the manual unchanged, add some text to the
description to make sure we've actually found an option. if the
manual changes this time, keep the converted description
this procedure converts 80% of nixos options to markdown. around 2000
options remain to be inspected, but most of those fail the "does not
change the manual output check": currently the MD conversion process
does not faithfully convert docbook tags like <code> and <package>, so
any option using such tags will not be converted at all.
If a host key file is a symlink pointing to an as of yet non-existent
file, we don't want to remove it, but instead follow the symlink and
create the file at that location.
See https://github.com/nix-community/impermanence/issues/101 for more
information on the issue the original behavior creates.
The group configuration parameter allow to share access to yggdrasil
control socket with the users in the system. In the version we propose,
it is null by default so that only root can access the control socket,
but let user create their own group if they need.
Remove User= durective in systemd unit. Should a user with the specified
name already exist in the system, it would be used silently instead of a
dynamic user which could be a security concern.
Since version 0.4 Yggdrasil works again using systemd's DynamicUser option.
This patch reenables it to improve security.
We tested this with both persistent and non-persistent keys. Everything
seems to work fine.
Suppose you want to provide a LDAP-based directory search to your
homeserver via a service-user with a bind-password. To make sure that
this doesn't end up in the Nix store, it's now possible to set a
substitute for the bindPassword like
services.mxisd.extraConfig.ldap.connection = {
# host, bindDn etc.
bindPassword = "$LDAP_BIND_PW";
};
and write the actual secret into an environment file that's readable for
`mxisd.service` containing
LDAP_BIND_PW=<your secret bind pw>
and the following setting in the Nix expression:
services.mxisd.environmentFile = "/runs/ecrets/mxisd";
(cherry picked from commit aa25ce7aa1a89618e4257fd46c7d20879f54c728)
...by using `replace-secret` instead of `sed` when injecting the
password into the ddclient config file. (Verified with `execsnoop`.)
Ref https://github.com/NixOS/nixpkgs/issues/156400.
Use `networking.resolvconf.package` to allow DNS entries to be set using
the system-wide resolver implementation instead of hardcoding systemd or
openresolv.
Extend the tests by adding DNS entries and making one of the peers use
systemd-networkd (hence systemd-resolved).
Also add a few `networkd`-specific settings.
A git command was failing in the test with
error: unable to get random bytes for temporary file: Operation not permitted
error: unable to create temporary file: Operation not permitted
error: .Radicale.lock: failed to insert into database
error: unable to index file '.Radicale.lock'
verbose is a debugging setting one step noisier than debug and should only be turned on when debugging because it leaks quite some credentials and tokens in the journalctl.
we expose it under settings instead of at the listener toplevel because
mosquitto seems to pick the addresses it will listen on
nondeterministically from the set of addresses configured on the
interface being bound to. encouraging its use by putting it into the
toplevel options for a listener seems inadvisable.
Introduced in OpenSSH 9.0 it became the part of the default kexAlgorithm
selection, visibile in sshd_config(5).
It is also enabled by default in the OpenSSH client, as can be seen from
$ ssh -Q KexAlgorithms
Also clarifies that we use the referenced documents as the lower bound,
given that they haven't been updated for 5-7y.
Tailscale uses policy routing to enable certain traffic to bypass
routes that lead into the Tailscale mesh. NixOS's reverse path
filtering setup doesn't understand the policy routing at play,
and so incorrectly interprets some of this traffic as spoofed.
Since this only breaks some features of Tailscale, merely warn
users about it, rather than make it a hard error.
Updates tailscale/tailscale#4432
Signed-off-by: David Anderson <dave@natulte.net>
For some features, tailscaled uses getent(1) to get the shell
of OS users. getent(1) is in the glibc derivation. Without this
derivation in the path, tailscale falls back to /bin/sh for all
users.
Signed-off-by: David Anderson <dave@natulte.net>
In a previous PR [1], the conditional to generate a new host key file
was changed to also include the case when the file exists, but has zero
size. This could occur when the system is uncleanly powered off shortly
after first boot.
However, ssh-keygen prompts the user before overwriting a file. For
example:
$ touch hi
$ ssh-keygen -f hi
Generating public/private rsa key pair.
hi already exists.
Overwrite (y/n)?
So, lets just try to remove the empty file (if it exists) before running
ssh-keygen.
[1] https://github.com/NixOS/nixpkgs/pull/141258
Without this, if the network goes down for a while, systemd will give up after 5 restarts:
Scheduled restart job, restart counter is at 5.
Stopped Nebula VPN service for myvpn.
nebula@myvpn.service: Start request repeated too quickly.
Failed with result 'exit-code'.
Failed to start Nebula VPN service for myvpn.
Most network services need this, but for VPNs it's extra important.
See https://blog.prosody.im/prosody-0.12.0-released for more
informations.
We remove the various lua wrappers introduced by
6799a91843 and
16d0b4a69f. It seems like we don't need
them anymore. I'm not brave enough to dig into the Lua machinery to
see what resolved that. Sorry, you'll have to trust me on that one.
We should probably think about the migration from http_upload to
http_file_share for the NixOS module. It's not trivial, we need to
make sure we don't break the already uploaded URLs.
This commit refactors the way how configuration files are deployed to
the `/etc/asterisk` directory.
The current solution builds a Nix derivation containing all config files
and symlinks it to `/etc/asterisk`. The problem with that approach is
that it is not possible to provide additional configuration that should
not be written to the Nix store, i.e. files containing credentials.
The proposed solution changes the creation of configuration files so
that each configuration file gets symlinked to `/etc/asterisk`
individually so that it becomes possible to provide additional config
files to `/etc/asterisk` as well.
This avoids the scenario where you activate a new config over Tailscale,
and a long delay between the "stop services" and "start services" phases
of the activation script lead to your terminal freezing for tens of
seconds, until tailscaled finally gets started again and the session
recovers.
Per the documentation of stopIfChanged, this is only safe to do if the
service definition is robust to stopping the old process using the new
service definition. As the maintainer of the upstream systemd unit, I
can confirm that Tailscale is robust to this scenario: it has to be
in order to work right on several other distros that just do
unpack-then-restart, rather than the more complex stop-unpack-start
dance.
Signed-off-by: David Anderson <dave@natulte.net>
1. Update the default values of several addresses-related settings
that have been changed by upstream.
2. Make `dns.address` take multiple addresses. This is needed
for dual stack, now working by default.
