d9396a82ce
Flake lock file updates: • Updated input 'nixpkgs': 'github:NixOS/nixpkgs/0d1c0755a8bce318ae64b95cdbd28377aa2edf29' (2023-06-19) → 'github:NixOS/nixpkgs/3fb3ce0b6b84d3b4e7b49e142da9c5764b563058' (2023-06-22) |
||
---|---|---|
.github | ||
src | ||
tests | ||
.gitignore | ||
.mergify.yml | ||
.nix-version | ||
.prettierrc.js | ||
default.nix | ||
flake.lock | ||
flake.nix | ||
LICENSE.md | ||
meson.build | ||
README.md | ||
shell.nix | ||
treefmt.toml |
nix-eval-jobs
This project evaluates nix attributes sets in parallel with streamable json output. This is useful for time and memory intensive evaluations such as NixOS machines, i.e. in a CI context. The evaluation is done with a controllable number of threads that are restarted when their memory consumption exceeds a certain threshold.
To facilitate integration, nix-eval-jobs creates garbage collection roots for each evaluated derivation (drv file, not the build) within the provided attribute. This prevents race conditions between the nix garbage collection service and user-started nix builds processes.
Why using nix-eval-jobs?
- Faster evaluation by using threads
- Memory used for evaluation is reclaimed after nix-eval-jobs finish, so that the build can use it.
- Evaluation of jobs can fail individually
Example
In the following example we evaluate the hydraJobs attribute of the patchelf flake:
$ nix-eval-jobs --gc-roots-dir gcroot --flake 'github:NixOS/patchelf#hydraJobs'
{"attr":"coverage","attrPath":["coverage"],"drvPath":"/nix/store/8hq9f09xa5s6g9m02lw0sw59kkkvj57c-patchelf-coverage-0.15.0.drv","name":"patchelf-coverage-0.15.0","outputs":{"out":"/nix/store/dwf255bdbfvvbiqak941r83zlvxyipcs-patchelf-coverage-0.15.0"},"system":"x86_64-linux"}
{"attr":"release","attrPath":["release"],"drvPath":"/nix/store/ip9dy4vlyha5a7kq4bnf4pxk0sfwjfda-patchelf-0.15.0.drv","name":"patchelf-0.15.0","outputs":{"out":"/nix/store/5z9ynn29asakf1b5736im2glcqpf6s2f-patchelf-0.15.0"},"system":"x86_64-linux"}
{"attr":"tarball","attrPath":["tarball"],"drvPath":"/nix/store/g1alnfi3mrkcb9blclr77fpyp35mpsdd-patchelf-tarball-0.15.0.drv","name":"patchelf-tarball-0.15.0","outputs":{"out":"/nix/store/iy0w42pffhjg6wy0w46r4cjc1yjk410y-patchelf-tarball-0.15.0"},"system":"x86_64-linux"}
The output here is newline-seperated json according to https://jsonlines.org.
The code is derived from hydra's eval-jobs executable.
Further options
$ nix-eval-jobs --help
USAGE: nix-eval-jobs [options] expr
--arg Pass the value *expr* as the argument *name* to Nix functions.
--argstr Pass the string *string* as the argument *name* to Nix functions.
--check-cache-status Check if the derivations are present locally or in any configured substituters (i.e. binary cache). The information will be exposed in the `isCached` field of the JSON output.
--debug Set the logging verbosity level to 'debug'.
--eval-store The Nix store to use for evaluations.
--expr treat the argument as a Nix expression
--flake build a flake
--gc-roots-dir garbage collector roots directory
--help show usage information
--impure allow impure expressions
--include Add *path* to the list of locations used to look up `<...>` file names.
--log-format Set the format of log output; one of `raw`, `internal-json`, `bar` or `bar-with-logs`.
--max-memory-size maximum evaluation memory size
--meta include derivation meta field in output
--option Set the Nix configuration setting *name* to *value* (overriding `nix.conf`).
--override-flake Override the flake registries, redirecting *original-ref* to *resolved-ref*.
--quiet Decrease the logging verbosity level.
--show-trace print out a stack trace in case of evaluation errors
--verbose Increase the logging verbosity level.
--workers number of evaluate workers
Potential use-cases for the tool
Faster evaluator in deployment tools. When evaluating NixOS machines, evaluation can take several minutes when run on a single core. This limits scalability for large deployments with deployment tools such as NixOps.
Faster evaluator in CIs. In addition to evaluation speed for CIs, it is also useful if evaluation of individual jobs in CIs can fail, as opposed to failing the entire jobset. For CIs that allow dynamic build steps to be created, one can also take advantage of the fact that nix-eval-jobs outputs the derivation path separately. This allows separate logs and success status per job instead of a single large log file. In the wiki we collect example ci configuration for various CIs.
Organisation of this repository
On the main
branch we target nixUnstable. When a release of nix happens, we
fork for a release branch i.e. release-2.8
and change the nix version in
.nix-version
. Changes and improvements made in main
also may be backported
to these release branches. At the time of writing we only intent to support the
latest release branch.
Projects using nix-eval-jobs
- colmena - A simple, stateless NixOS deployment tool
- robotnix - Build Android (AOSP) using Nix, used in their CI
FAQ
nix-eval-jobs consumes too much memory / is too slow
By default, nix-eval-jobs spawns as many worker processes as there are hardware threads in the system and limits the memory usage for each worker to 4GB.
However, keep in mind that each worker process may need to re-evaluate shared dependencies of the attributes, which can introduce some overhead for each evaluation or cause workers to exceed their memory limit. If you encounter these situations, you can tune the following options:
--workers
: This option allows you to set the number of evaluation workers that
nix-eval-jobs should spawn. You can increase or decrease this number to
optimize the evaluation speed and memory usage. For example, if you have a
system with many CPU cores but limited memory, you may want to reduce the
number of workers to avoid exceeding the memory limit.
--max-memory-size
: This option allows you to adjust the memory limit for each
worker process. By default, it's set to 4GiB, but you can increase or decrease
this value as needed. For example, if you have a system with a lot of memory
and want to speed up the evaluation, you may want to increase the memory limit
to allow workers to cache more data in memory before getting restarted by
nix-eval-jobs.
Overall, tuning these options can help you optimize the performance and memory usage of nix-eval-jobs to better fit your system and evaluation needs.