Since cd1dedac67 systemd-networkd has it's
netlink socket created via a systemd.socket unit. One might think that
this doesn't make much sense since networkd is just going to create it's
own socket on startup anyway. The difference here is that we have
configuration-time control over things like socket buffer sizes vs
compile-time constants.
For larger setups where networkd has to create a lot of (virtual)
devices the default buffer size of currently 128MB is not enough.
A good example is a machine with >100 virtual interfaces (e.g.,
wireguard tunnels, VLANs, …) that all have to be brought up during
startup. The receive buffer size will spike due to all the generated
message from the new interfaces. Eventually some of the message will be
dropped since there is not enough (permitted) buffer space available.
By having networkd start through / with a netlink socket created by
systemd we can configure the `ReceiveBufferSize` parameter in the socket
options without recompiling networkd.
Since the actual memory requirements depend on hardware, timing, exact
configurations etc. it isn't currently possible to infer a good default
from within the NixOS module system. Administrators are advised to
monitor the logs of systemd-networkd for `rtnl: kernel receive buffer
overrun` spam and increase the memory as required.
Note: Increasing the ReceiveBufferSize doesn't allocate any memory. It
just increases the upper bound on the kernel side. The memory allocation
depends on the amount of messages that are queued on the kernel side of
the netlink socket.
udev gained native support to handle FIDO security tokens, so we don't
need a module which only added the now obsolete udev rules.
Fixes: https://github.com/NixOS/nixpkgs/issues/76482
This fixes the output of "hostname --fqdn" (previously the domain name
was not appended). Additionally it's now possible to use the FQDN.
This works by unconditionally adding two entries to /etc/hosts:
127.0.0.1 localhost
::1 localhost
These are the first two entries and therefore gethostbyaddr() will
always resolve "127.0.0.1" and "::1" back to "localhost" [0].
This works because nscd (or rather the nss-files module) returns the
first matching row from /etc/hosts (and ignores the rest).
The FQDN and hostname entries are appended later to /etc/hosts, e.g.:
127.0.0.2 nixos-unstable.test.tld nixos-unstable
::1 nixos-unstable.test.tld nixos-unstable
Note: We use 127.0.0.2 here to follow nss-myhostname (systemd) as close
as possible. This has the advantage that 127.0.0.2 can be resolved back
to the FQDN but also the drawback that applications that only listen to
127.0.0.1 (and not additionally ::1) cannot be reached via the FQDN.
If you would like this to work you can use the following configuration:
```nix
networking.hosts."127.0.0.1" = [
"${config.networking.hostName}.${config.networking.domain}"
config.networking.hostName
];
```
Therefore gethostbyname() resolves "nixos-unstable" to the FQDN
(canonical name): "nixos-unstable.test.tld".
Advantages over the previous behaviour:
- The FQDN will now also be resolved correctly (the entry was missing).
- E.g. the command "hostname --fqdn" will now work as expected.
Drawbacks:
- Overrides entries form the DNS (an issue if e.g. $FQDN should resolve
to the public IP address instead of 127.0.0.1)
- Note: This was already partly an issue as there's an entry for
$HOSTNAME (without the domain part) that resolves to
127.0.1.1 (!= 127.0.0.1).
- Unknown (could potentially cause other unexpected issues, but special
care was taken).
[0]: Some applications do apparently depend on this behaviour (see
c578924) and this is typically the expected behaviour.
Co-authored-by: Florian Klink <flokli@flokli.de>
The `network-link-${i.name}` units raced with other things trying to
configure the interface, or ran before the interface was available.
Instead of running our own set of shell scripts on boot, and hoping
they're executed at the right time, we can make use of udev to configure
the interface *while they appear*, by providing `.link` files in
/etc/systemd/network/*.link to set MACAddress and MTUBytes.
This doesn't require networkd to be enabled, and is populated properly
on non-networkd systems since
https://github.com/NixOS/nixpkgs/pull/82941.
This continues clean-up work done in
https://github.com/NixOS/nixpkgs/pull/85170 for the scripted networking
stack.
The only leftover part of the `network-link-${i.name}` unit (bringing
the interface up) is moved to the beginning of the
`network-addresses-${i.name}` unit.
Fixes: https://github.com/NixOS/nixpkgs/issues/74471
Closes: https://github.com/NixOS/nixpkgs/pull/87116
I hate the thing too even though I made it, and rather just get rid of
it. But we can't do that yet. In the meantime, this brings us more
inline with autoconf and will make it slightly easier for me to write a
pkg-config wrapper, which we need.
'nix build' is an experimental command so we shouldn't use it
yet. (nixos-rebuild also uses 'nix', but only when using flakes, which
are themselves an experimental feature.)
This follows upstreams change in documentation. While the `[DHCP]`
section might still work it is undocumented and we should probably not
be using it anymore. Users can just upgrade to the new option without
much hassle.
I had to create a bit of custom module deprecation code since the usual
approach doesn't support wildcards in the path.
Systemd upstream has deprecated CriticalConnection with v244 in favor of
KeepConnection as that seems to be more flexible:
The CriticalConnection= setting in .network files is now deprecated,
and replaced by a new KeepConfiguration= setting which allows more
detailed configuration of the IP configuration to keep in place.
Rework withExtensions / buildEnv to handle currently enabled
extensions better and make them compatible with override. They now
accept a function with the named arguments enabled and all, where
enabled is a list of currently enabled extensions and all is the set
of all extensions. This gives us several nice properties:
- You always get the right version of the list of currently enabled
extensions
- Invocations chain
- It works well with overridden PHP packages - you always get the
correct versions of extensions
As a contrived example of what's possible, you can add ImageMagick,
then override the version and disable fpm, then disable cgi, and
lastly remove the zip extension like this:
{ pkgs ? (import <nixpkgs>) {} }:
with pkgs;
let
phpWithImagick = php74.withExtensions ({ all, enabled }: enabled ++ [ all.imagick ]);
phpWithImagickWithoutFpm743 = phpWithImagick.override {
version = "7.4.3";
sha256 = "wVF7pJV4+y3MZMc6Ptx21PxQfEp6xjmYFYTMfTtMbRQ=";
fpmSupport = false;
};
phpWithImagickWithoutFpmZip743 = phpWithImagickWithoutFpm743.withExtensions (
{ enabled, all }:
lib.filter (e: e != all.zip) enabled);
phpWithImagickWithoutFpmZipCgi743 = phpWithImagickWithoutFpmZip743.override {
cgiSupport = false;
};
in
phpWithImagickWithoutFpmZipCgi743
It currently says that everything will be backward compatible between lego and simp-le certificates, but it’s not.
(cherry picked from commit 21c4a33ceef77dec2b821f7164e13971862d5575)
According to my analysis the last critical fix went into v5.4.23, I have
confirmed this by running WebGL over night and haven't seen a single
i915 GPU hang. Lets remove the notes from the release notes.
(cherry picked from commit da764d22ce3b698707861d58824843ded87cbb0a)
This is an backward incompatible change from upstream dhcpcd [0], as
this could have easily locked me out of my box.
As dhcpcd doesn't allow to use only a blacklist (denyinterfaces in
dhcpcd.conf) of devices and use all remaining devices, while explicitly
allowing some interfaces like bridges, I think the best option would be
to not change anything about it and just educate the users here about
that edge case and how to solve it.
[0] https://roy.marples.name/archives/dhcpcd-discuss/0002621.html
(cherry picked from commit eeeb2bf8035b309a636d596de6a3b1d52ca427b1)
Also removed `pkgs.hydra-flakes` since flake-support has been merged
into master[1]. Because of that, `pkgs.hydra-unstable` is now compiled
against `pkgs.nixFlakes` and currently requires a patch since Hydra's
master doesn't compile[2] atm.
[1] https://github.com/NixOS/hydra/pull/730
[2] https://github.com/NixOS/hydra/pull/732
Some changes might require manual migration steps:
"Due to changes to the way in which Gollum handles filenames, you may
have to change some links in your wiki when migrating from gollum 4.x.
See the release notes [0] for more details. You may find the
bin/gollum-migrate-tags script helpful to accomplish this. Also see the
--lenient-tag-lookup option for making tag lookup backwards compatible
with 4.x, though note that this will decrease performance on large wikis
with many tags." (source: [1])
[0]: https://github.com/gollum/gollum/wiki/5.0-release-notes
[1]: https://github.com/gollum/gollum/blob/v5.0.0/HISTORY.md
So now we have only packages for human interaction in php.packages and
only extensions in php.extensions. With this php.packages.exts have
been merged into the same attribute set as all the other extensions to
make it flat and nice.
The nextcloud module have been updated to reflect this change as well
as the documentation.
Upgrades Hydra to the latest master/flake branch. To perform this
upgrade, it's needed to do a non-trivial db-migration which provides a
massive performance-improvement[1].
The basic ideas behind multi-step upgrades of services between NixOS versions
have been gathered already[2]. For further context it's recommended to
read this first.
Basically, the following steps are needed:
* Upgrade to a non-breaking version of Hydra with the db-changes
(columns are still nullable here). If `system.stateVersion` is set to
something older than 20.03, the package will be selected
automatically, otherwise `pkgs.hydra-migration` needs to be used.
* Run `hydra-backfill-ids` on the server.
* Deploy either `pkgs.hydra-unstable` (for Hydra master) or
`pkgs.hydra-flakes` (for flakes-support) to activate the optimization.
The steps are also documented in the release-notes and in the module
using `warnings`.
`pkgs.hydra` has been removed as latest Hydra doesn't compile with
`pkgs.nixStable` and to ensure a graceful migration using the newly
introduced packages.
To verify the approach, a simple vm-test has been added which verifies
the migration steps.
[1] https://github.com/NixOS/hydra/pull/711
[2] https://github.com/NixOS/nixpkgs/pull/82353#issuecomment-598269471
It's impossible to move two major-versions forward when upgrading
Nextcloud. This is an issue when comming from 19.09 (using Nextcloud 16)
and trying to upgrade to 20.03 (using Nextcloud 18 by default).
This patch implements the measurements discussed in #82056 and #82353 to
improve the update process and to circumvent similar issues in the
future:
* `pkgs.nextcloud` has been removed in favor of versioned attributes
(currently `pkgs.nextcloud17` and `pkgs.nextcloud18`). With that
approach we can safely backport major-releases in the future to
simplify those upgrade-paths and we can select one of the
major-releases as default depending on the configuration (helpful to
decide whether e.g. `pkgs.nextcloud17` or `pkgs.nextcloud18` should be
used on 20.03 and `master` atm).
* If `system.stateVersion` is older than `20.03`, `nextcloud17` will be
used (which is one major-release behind v16 from 19.09). When using a
package older than the latest major-release available (currently v18),
the evaluation will cause a warning which describes the issue and
suggests next steps.
To make those package-selections easier, a new option to define the
package to be used for the service (namely
`services.nextcloud.package`) was introduced.
* If `pkgs.nextcloud` exists (e.g. due to an overlay which was used to
provide more recent Nextcloud versions on older NixOS-releases), an
evaluation error will be thrown by default: this is to make sure that
`services.nextcloud.package` doesn't use an older version by accident
after checking the state-version. If `pkgs.nextcloud` is added
manually, it needs to be declared explicitly in
`services.nextcloud.package`.
* The `nixos/nextcloud`-documentation contains a
"Maintainer information"-chapter which describes how to roll out new
Nextcloud releases and how to deal with old (and probably unsafe)
versions.
Closes#82056
Dropbear lags behind OpenSSH significantly in both support for modern
key formats like `ssh-ed25519`, let alone the recently-introduced
U2F/FIDO2-based `sk-ssh-ed25519@openssh.com` (as I found when I switched
my `authorizedKeys` over to it and promptly locked myself out of my
server's initrd SSH, breaking reboots), as well as security features
like multiprocess isolation. Using the same SSH daemon for stage-1 and
the main system ensures key formats will always remain compatible, as
well as more conveniently allowing the sharing of configuration and
host keys.
The main reason to use Dropbear over OpenSSH would be initrd space
concerns, but NixOS initrds are already large (17 MiB currently on my
server), and the size difference between the two isn't huge (the test's
initrd goes from 9.7 MiB to 12 MiB with this change). If the size is
still a problem, then it would be easy to shrink sshd down to a few
hundred kilobytes by using an initrd-specific build that uses musl and
disables things like Kerberos support.
This passes the test and works on my server, but more rigorous testing
and review from people who use initrd SSH would be appreciated!
Running the manual on a TTY is useless in the graphical ISOs and not
particularly useful in non-graphical ISOs (since you can also run
'nixos-help').
Fixes#83157.
This mirrors the behaviour of systemd - It's udev that parses `.link`
files, not `systemd-networkd`.
This was originally applied in 36ef112a47,
but was reverted due to 1115959a8d causing
evaluation errors on hydra.
* Linkify all service options used in the code-examples.
* Demonstrated the use of `riot-web.override {}`.
* Moved the example how to configure a postgresql-database for
`matrix-synapse` to this document from the 20.03 release-notes.
...even when networkd is disabled
This reverts commit ce78f3ac70, reversing
changes made to dc34da0755.
I'm sorry; Hydra has been unable to evaluate, always returning
> error: unexpected EOF reading a line
and I've been unable to reproduce the problem locally. Bisecting
pointed to this merge, but I still can't see what exactly was wrong.
