This project aims to provide a lean configuration for users that want to run lancache on their nix machines without having to resort to running a docker container.
It is designed to run alongside your configurations while keeping feature-parity with the upstream project.
The configurations are almost 1:1 with lancachenet/monolithic and all credit for the nginx configurations go to the lancache authors.
services.lancache = {
enable = true;
cacheLocation = "/mnt/storage/lancache";
logPrefix = "/var/log/nginx/lancache";
# the listen address you want lancache to use.
listenAddress = "10.0.0.2";
# optional options with their defaults
/*
The upstream DNS servers the cache should use. The defaults are
Cloudflare's DNS.
Corresponds to the environment variable: `UPSTREAM_DNS` in standard
lancache config.
*/
upstreamDns = [
"1.1.1.1"
"1.0.0.1"
];
/*
The amount of disk space the container should use for caching data.
Specified in gigabytes.
Corresponds to the environment variable: `CACHE_DISK_SIZE` in
standard lancache config.
*/
cacheDiskSize = "1000g";
/*
Amount of index memory for the nginx cache manager. Lancache team
recommends `250m` of index memory per `1TB` of `cacheDiskSize`
Corresponds to the environment variable: `CACHE_INDEX_SIZE` in
standard lancache config.
*/
cacheIndexSize = "500m";
/*
The maximum amount of time a file should be held in cache. There is
usually no reason to reduce this - the cache will automatically
remove the oldest content if it needs the space.
Corresponds to the environment variable: `CACHE_MAX_AGE` in standard
lancache config.
*/
cacheMaxAge = "3560d";
/*
Sets the minimum free disk space that must be kept at all times. When the
available free space drops below the set amount for any reason, the cache
server will begin pruning content to free up space. Specified in gigabytes.
Corresponds to the environment variable: `MIN_FREE_DISK` in standard
lancache config.
*/
minFreeDisk = "10g";
/*
See the guide before changing this. It WILL invalidate any currently cached data.
link: https://lancache.net/docs/advanced/tuning-cache/#b---tweaking-slice-size
Corresponds to the environment variable: `CACHE_SLICE_SIZE` in standard
lancache config.
*/
sliceSize = "1m";
/*
Sets the default logging format
Nothing corresponds to this option in the standard lancache config
Set it to cachelog-json to output json logs
*/
logFormat = "cachelog";
domainsPackage = pkgs.fetchFromGitHub {
owner = "uklans";
repo = "cache-domains";
rev = "1f5897f4dacf3dab5f4d6fca2fe497d3327eaea9";
sha256 = "sha256-xrHuYIrGSzsPtqErREMZ8geawvtYcW6h2GyeGMw1I88=";
};
# Defines the number of worker processes.
workerProcesses = "auto";
};These two are the major services that should be exposed:
- the catch-all http server that you can point the dns entries of cachable cdns to
- a transparent HTTPS proxy listening on
listenAddress:443
One thing to note is that if you are listening to 0.0.0.0 on any of your
virtual hosts, it will clash with these configurations. The nixos nginx module
uses 0.0.0.0 as the defaultListenAddress so I recommend you change that
services.nginx.defaultListenAddresses = [ "192.168.100.2" ];You might see errors like this:
nginx: [emerg] bind() to [::]:80 failed (98: Address already in use)
nginx: [emerg] bind() to [::]:443 failed (98: Address already in use)
This means you have vhosts already listening on 443/80. You should assign
an extra ip address to your interface to get around this and use that ip
address for lancache.
Setting the defaultListenAddress (see above) is important too if you want your vhosts to just work™ without specifying listen addresses.
networking.interfaces.${interface}.ipv4.addresses = [
{
address = "192.168.100.3";
prefixLength = 24;
}
];
services.lancache.listenAddress = "192.168.100.3";