Revise or remove?
No matter how well it's written, this basically does what the instanced services method does, but users make the individual socket and service files by hand. Even if individual resolvers need different configurations, that could be achieved by overriding the instanced services. Some of this I would recycle into the instanced services section (Tip, Lastly). quequotion (talk) 16:04, 25 January 2017 (UTC)
Redundant DNSCrypt providers
To use additional dnscrypt providers, copy
/usr/lib/systemd/system/dnscrypt-proxy.service to a new file,
/etc/systemd/system/dnscrypt-proxy-short-name.here.service and specify a different resolver using the
-R flag with a short name from
[Service] ExecStart= ExecStart=/usr/bin/dnscrypt-proxy -R short-name.here
/usr/lib/systemd/system/dnscrypt-proxy.socket to a new file,
/etc/systemd/system/dnscrypt-proxy-short-name.here.socket and, specify another port.
Lastly, update your local DNS cache program to point to new service's port. For example, with unbound the configuration file would look like if using ports
5353 for the original socket and
5354 for the new socket.
Backup DNSCrypt resolver - especially with the new configuration file
Usually when setting a dns resolver you will always have the option to set a second/backup dns resolver (android,windows,networkmanager,router, what ever).
I think the wiki should cover a way on how to achieve the same with dnscrypt. Especially as some if the dnscrypt resolvers like to go offline every now and then (looking at you dnscrypt.eu-nl).
I have a running setup (which caused me some struggles to achieve that setup) but I have no idea how to replicate it. Espcially with the new configuration file which seems like it will only cover one dnscrypt instance?
Right now I have 2x dnscrypt running in systemd and the resolver.conf will choose which ever is online/working.
- It already describes that: DNSCrypt#Redundant_DNSCrypt_providers -- Lahwaacz (talk) 12:02, 30 December 2016 (UTC)