systemd 258 has the following changes noted in systemd.resource-control(5):
> `CPUAccounting=` setting is deprecated, because it is always available on the unified cgroup hierarchy and such setting has no effect.
This commit removes it from the three services using it directly, as well as one instance of example text.
Signed-off-by: benaryorg <binary@benary.org>
One of the nicest perks of using scx is the ability to easily switch
between schedulers at run-time, but this is currently not possible in
NixOS, given that the scx service only uses the scheduler from the
config file.
It's possible, however, to override the scheduler and its flags using
the `SCX_SCHEDULER_OVERRIDE` and `SCX_FLAGS_OVERRIDE` environment
variables, respectively (see [more
details](https://en.opensuse.org/Pluggable_CPU_schedulers#Temporarily_switch_to_a_different_scheduler)).
So the solution here is to change the scx service executable and pass
those env vars to the service at run-time falling back to the values
from the module config as defaults.
After this change, it would be possible to override schedulers like so:
```shellSession
$ sudo systemctl set-environment SCX_SCHEDULER_OVERRIDE='scx_lavd'
$ sudo systemctl set-environment SCX_FLAGS_OVERRIDE='--performance'
$ sudo systemctl restart scx.service
```
```shellSession
$ journalctl -u scx.service -b 0 -e
Oct 18 13:09:25 nixos systemd[1]: Started SCX scheduler daemon.
Oct 18 13:09:25 nixos bash[1829848]: 13:09:25 [INFO] Performance mode is
enabled.
Oct 18 13:09:25 nixos bash[1829848]: 13:09:25 [INFO] Energy model won't
be used for CPU preference order.
Oct 18 13:09:26 nixos bash[1829848]: 13:09:26 [WARN] libbpf: map
'lavd_ops': BPF map skeleton link is uninitialized
Oct 18 13:09:26 nixos bash[1829848]: 13:09:26 [INFO] scx_lavd scheduler
is initialized (build ID: 1.0.16 x86_64-unknown-linux-gnu)
Oct 18 13:09:26 nixos bash[1829848]: 13:09:26 [INFO] scx_lavd scheduler
starts running.
```
And also restore to the module defaults:
```shellSession
$ sudo systemctl unset-environment SCX_SCHEDULER_OVERRIDE
$ sudo systemctl unset-environment SCX_FLAGS_OVERRIDE
$ sudo systemctl restart scx.service
```
```shellSession
$ journalctl -u scx.service -b 0 -e
Oct 18 13:12:56 nixos systemd[1]: Stopped SCX scheduler daemon.
Oct 18 13:12:56 nixos systemd[1]: Started SCX scheduler daemon.
Oct 18 13:12:56 nixos bash[1833220]: 13:12:56 [INFO] NUMA nodes: 1
Oct 18 13:12:56 nixos bash[1833220]: 13:12:56 [INFO] Disabling NUMA
optimizations
Oct 18 13:12:56 nixos bash[1833220]: 13:12:56 [INFO] scx_bpfland 1.0.16
x86_64-unknown-linux-gnu SMT on
```
prefect: add dburl to worker
prefect: use same state directory
prefect: fix worker environment
prefect: create user
prefect: use datadir for sqlite url
prefect: make datadir writable
prefect: don't protect home
prefect fix sqlite url
prefect: fix state directory
prefect: user should not be systemuser
prefect: set to normal user
add prefect to systempackages
try user with same name
prefect use prefect_home
do not set database url
revert to dynamic user
prefect: add tests
prefect: fix port to string
Fixes the following warning when setting
`systemd.enableStrictShellChecks = true`:
> SC2174 (warning): When used with -p, -m only applies to the deepest
> directory.
Since it could conceivably be possible for `/var` to not exist when
running this script, the fix also explicitly tries to create /var with
the normal 0755 permissions.
After final improvements to the official formatter implementation,
this commit now performs the first treewide reformat of Nix files using it.
This is part of the implementation of RFC 166.
Only "inactive" files are reformatted, meaning only files that
aren't being touched by any PR with activity in the past 2 months.
This is to avoid conflicts for PRs that might soon be merged.
Later we can do a full treewide reformat to get the rest,
which should not cause as many conflicts.
A CI check has already been running for some time to ensure that new and
already-formatted files are formatted, so the files being reformatted here
should also stay formatted.
This commit was automatically created and can be verified using
nix-build a08b3a4d19.tar.gz \
--argstr baseRev b32a094368
result/bin/apply-formatting $NIXPKGS_PATH
This adds a `services.scx.enable` option to enable sched-ext schedulers.
Requires a kernel with sched-ext enabled (6.12+) or a kernel with the patchset.
requiredKernelConfigs are taken from https://cateee.net/lkddb/web-lkddb/SCHED_CLASS_EXT.html
these changes were generated with nixq 0.0.2, by running
nixq ">> lib.mdDoc[remove] Argument[keep]" --batchmode nixos/**.nix
nixq ">> mdDoc[remove] Argument[keep]" --batchmode nixos/**.nix
nixq ">> Inherit >> mdDoc[remove]" --batchmode nixos/**.nix
two mentions of the mdDoc function remain in nixos/, both of which
are inside of comments.
Since lib.mdDoc is already defined as just id, this commit is a no-op as
far as Nix (and the built manual) is concerned.
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.
Since https://github.com/NixOS/nixpkgs/pull/61321, local-fs.target is
part of sysinit.target again, meaning units without
DefaultDependencies=no will automatically depend on it, and the manual
set dependencies can be dropped.
Without fcrontab being setuid, every attempt by an user in the fcron
group to edit their own crontab (via `fcrontab -e`) results in the
following error:
```
2018-05-06 11:29:07 ERROR could not change euid to 273: Operation not permitted
2018-05-06 11:29:07 ERROR fcron child aborted: this does not affect the main fcron daemon, but this may prevent a job from being run or an email from being sent.
```
Adding setuid by hand has resolved this issue and aligns with the way
fcrontab is installed on other distributions.
"batch" is a shell script so invoking it via setuid wrapper never worked
anyway. (The kernel drops perms on executables with shebang.) A previous
nixpkgs commit made "batch" invoke the NixOS setuid "at" wrapper to gain
needed privileges.
Thanks to @yesbox for noticing.
