nixpkgs/nixos/tests/acme.nix

240 lines
9.4 KiB
Nix
Raw Normal View History

let
commonConfig = ./common/acme/client;
dnsScript = {writeScript, dnsAddress, bash, curl}: writeScript "dns-hook.sh" ''
#!${bash}/bin/bash
set -euo pipefail
echo '[INFO]' "[$2]" 'dns-hook.sh' $*
if [ "$1" = "present" ]; then
${curl}/bin/curl --data '{"host": "'"$2"'", "value": "'"$3"'"}' http://${dnsAddress}:8055/set-txt
else
${curl}/bin/curl --data '{"host": "'"$2"'"}' http://${dnsAddress}:8055/clear-txt
fi
'';
in import ./make-test-python.nix ({ lib, ... }: {
name = "acme";
meta.maintainers = lib.teams.acme.members;
Fix letsencrypt (#60219) * nixos/acme: Fix ordering of cert requests When subsequent certificates would be added, they would not wake up nginx correctly due to target units only being triggered once. We now added more fine-grained systemd dependencies to make sure nginx always is aware of new certificates and doesn't restart too early resulting in a crash. Furthermore, the acme module has been refactored. Mostly to get rid of the deprecated PermissionStartOnly systemd options which were deprecated. Below is a summary of changes made. * Use SERVICE_RESULT to determine status This was added in systemd v232. we don't have to keep track of the EXITCODE ourselves anymore. * Add regression test for requesting mutliple domains * Deprecate 'directory' option We now use systemd's StateDirectory option to manage create and permissions of the acme state directory. * The webroot is created using a systemd.tmpfiles.rules rule instead of the preStart script. * Depend on certs directly By getting rid of the target units, we make sure ordering is correct in the case that you add new certs after already having deployed some. Reason it broke before: acme-certificates.target would be in active state, and if you then add a new cert, it would still be active and hence nginx would restart without even requesting a new cert. Not good! We make the dependencies more fine-grained now. this should fix that * Remove activationDelay option It complicated the code a lot, and is rather arbitrary. What if your activation script takes more than activationDelay seconds? Instead, one should use systemd dependencies to make sure some action happens before setting the certificate live. e.g. If you want to wait until your cert is published in DNS DANE / TLSA, you could create a unit that blocks until it appears in DNS: ``` RequiredBy=acme-${cert}.service After=acme-${cert}.service ExecStart=publish-wait-for-dns-script ```
2019-08-29 14:32:59 +00:00
nodes = rec {
acme = { nodes, lib, ... }: {
imports = [ ./common/acme/server ];
networking.nameservers = lib.mkForce [
nodes.dnsserver.config.networking.primaryIPAddress
];
};
dnsserver = { nodes, pkgs, ... }: {
networking.firewall.allowedTCPPorts = [ 8055 53 ];
networking.firewall.allowedUDPPorts = [ 53 ];
systemd.services.pebble-challtestsrv = {
enable = true;
description = "Pebble ACME challenge test server";
wantedBy = [ "network.target" ];
serviceConfig = {
ExecStart = "${pkgs.pebble}/bin/pebble-challtestsrv -dns01 ':53' -defaultIPv6 '' -defaultIPv4 '${nodes.webserver.config.networking.primaryIPAddress}'";
# Required to bind on privileged ports.
AmbientCapabilities = [ "CAP_NET_BIND_SERVICE" ];
};
};
};
acmeStandalone = { nodes, lib, config, pkgs, ... }: {
Fix letsencrypt (#60219) * nixos/acme: Fix ordering of cert requests When subsequent certificates would be added, they would not wake up nginx correctly due to target units only being triggered once. We now added more fine-grained systemd dependencies to make sure nginx always is aware of new certificates and doesn't restart too early resulting in a crash. Furthermore, the acme module has been refactored. Mostly to get rid of the deprecated PermissionStartOnly systemd options which were deprecated. Below is a summary of changes made. * Use SERVICE_RESULT to determine status This was added in systemd v232. we don't have to keep track of the EXITCODE ourselves anymore. * Add regression test for requesting mutliple domains * Deprecate 'directory' option We now use systemd's StateDirectory option to manage create and permissions of the acme state directory. * The webroot is created using a systemd.tmpfiles.rules rule instead of the preStart script. * Depend on certs directly By getting rid of the target units, we make sure ordering is correct in the case that you add new certs after already having deployed some. Reason it broke before: acme-certificates.target would be in active state, and if you then add a new cert, it would still be active and hence nginx would restart without even requesting a new cert. Not good! We make the dependencies more fine-grained now. this should fix that * Remove activationDelay option It complicated the code a lot, and is rather arbitrary. What if your activation script takes more than activationDelay seconds? Instead, one should use systemd dependencies to make sure some action happens before setting the certificate live. e.g. If you want to wait until your cert is published in DNS DANE / TLSA, you could create a unit that blocks until it appears in DNS: ``` RequiredBy=acme-${cert}.service After=acme-${cert}.service ExecStart=publish-wait-for-dns-script ```
2019-08-29 14:32:59 +00:00
imports = [ commonConfig ];
networking.nameservers = lib.mkForce [
nodes.dnsserver.config.networking.primaryIPAddress
];
Fix letsencrypt (#60219) * nixos/acme: Fix ordering of cert requests When subsequent certificates would be added, they would not wake up nginx correctly due to target units only being triggered once. We now added more fine-grained systemd dependencies to make sure nginx always is aware of new certificates and doesn't restart too early resulting in a crash. Furthermore, the acme module has been refactored. Mostly to get rid of the deprecated PermissionStartOnly systemd options which were deprecated. Below is a summary of changes made. * Use SERVICE_RESULT to determine status This was added in systemd v232. we don't have to keep track of the EXITCODE ourselves anymore. * Add regression test for requesting mutliple domains * Deprecate 'directory' option We now use systemd's StateDirectory option to manage create and permissions of the acme state directory. * The webroot is created using a systemd.tmpfiles.rules rule instead of the preStart script. * Depend on certs directly By getting rid of the target units, we make sure ordering is correct in the case that you add new certs after already having deployed some. Reason it broke before: acme-certificates.target would be in active state, and if you then add a new cert, it would still be active and hence nginx would restart without even requesting a new cert. Not good! We make the dependencies more fine-grained now. this should fix that * Remove activationDelay option It complicated the code a lot, and is rather arbitrary. What if your activation script takes more than activationDelay seconds? Instead, one should use systemd dependencies to make sure some action happens before setting the certificate live. e.g. If you want to wait until your cert is published in DNS DANE / TLSA, you could create a unit that blocks until it appears in DNS: ``` RequiredBy=acme-${cert}.service After=acme-${cert}.service ExecStart=publish-wait-for-dns-script ```
2019-08-29 14:32:59 +00:00
networking.firewall.allowedTCPPorts = [ 80 ];
security.acme.certs."standalone.test" = {
webroot = "/var/lib/acme/acme-challenges";
Fix letsencrypt (#60219) * nixos/acme: Fix ordering of cert requests When subsequent certificates would be added, they would not wake up nginx correctly due to target units only being triggered once. We now added more fine-grained systemd dependencies to make sure nginx always is aware of new certificates and doesn't restart too early resulting in a crash. Furthermore, the acme module has been refactored. Mostly to get rid of the deprecated PermissionStartOnly systemd options which were deprecated. Below is a summary of changes made. * Use SERVICE_RESULT to determine status This was added in systemd v232. we don't have to keep track of the EXITCODE ourselves anymore. * Add regression test for requesting mutliple domains * Deprecate 'directory' option We now use systemd's StateDirectory option to manage create and permissions of the acme state directory. * The webroot is created using a systemd.tmpfiles.rules rule instead of the preStart script. * Depend on certs directly By getting rid of the target units, we make sure ordering is correct in the case that you add new certs after already having deployed some. Reason it broke before: acme-certificates.target would be in active state, and if you then add a new cert, it would still be active and hence nginx would restart without even requesting a new cert. Not good! We make the dependencies more fine-grained now. this should fix that * Remove activationDelay option It complicated the code a lot, and is rather arbitrary. What if your activation script takes more than activationDelay seconds? Instead, one should use systemd dependencies to make sure some action happens before setting the certificate live. e.g. If you want to wait until your cert is published in DNS DANE / TLSA, you could create a unit that blocks until it appears in DNS: ``` RequiredBy=acme-${cert}.service After=acme-${cert}.service ExecStart=publish-wait-for-dns-script ```
2019-08-29 14:32:59 +00:00
};
systemd.targets."acme-finished-standalone.test" = {
after = [ "acme-standalone.test.service" ];
wantedBy = [ "acme-standalone.test.service" ];
Fix letsencrypt (#60219) * nixos/acme: Fix ordering of cert requests When subsequent certificates would be added, they would not wake up nginx correctly due to target units only being triggered once. We now added more fine-grained systemd dependencies to make sure nginx always is aware of new certificates and doesn't restart too early resulting in a crash. Furthermore, the acme module has been refactored. Mostly to get rid of the deprecated PermissionStartOnly systemd options which were deprecated. Below is a summary of changes made. * Use SERVICE_RESULT to determine status This was added in systemd v232. we don't have to keep track of the EXITCODE ourselves anymore. * Add regression test for requesting mutliple domains * Deprecate 'directory' option We now use systemd's StateDirectory option to manage create and permissions of the acme state directory. * The webroot is created using a systemd.tmpfiles.rules rule instead of the preStart script. * Depend on certs directly By getting rid of the target units, we make sure ordering is correct in the case that you add new certs after already having deployed some. Reason it broke before: acme-certificates.target would be in active state, and if you then add a new cert, it would still be active and hence nginx would restart without even requesting a new cert. Not good! We make the dependencies more fine-grained now. this should fix that * Remove activationDelay option It complicated the code a lot, and is rather arbitrary. What if your activation script takes more than activationDelay seconds? Instead, one should use systemd dependencies to make sure some action happens before setting the certificate live. e.g. If you want to wait until your cert is published in DNS DANE / TLSA, you could create a unit that blocks until it appears in DNS: ``` RequiredBy=acme-${cert}.service After=acme-${cert}.service ExecStart=publish-wait-for-dns-script ```
2019-08-29 14:32:59 +00:00
};
services.nginx.enable = true;
services.nginx.virtualHosts."standalone.test" = {
Fix letsencrypt (#60219) * nixos/acme: Fix ordering of cert requests When subsequent certificates would be added, they would not wake up nginx correctly due to target units only being triggered once. We now added more fine-grained systemd dependencies to make sure nginx always is aware of new certificates and doesn't restart too early resulting in a crash. Furthermore, the acme module has been refactored. Mostly to get rid of the deprecated PermissionStartOnly systemd options which were deprecated. Below is a summary of changes made. * Use SERVICE_RESULT to determine status This was added in systemd v232. we don't have to keep track of the EXITCODE ourselves anymore. * Add regression test for requesting mutliple domains * Deprecate 'directory' option We now use systemd's StateDirectory option to manage create and permissions of the acme state directory. * The webroot is created using a systemd.tmpfiles.rules rule instead of the preStart script. * Depend on certs directly By getting rid of the target units, we make sure ordering is correct in the case that you add new certs after already having deployed some. Reason it broke before: acme-certificates.target would be in active state, and if you then add a new cert, it would still be active and hence nginx would restart without even requesting a new cert. Not good! We make the dependencies more fine-grained now. this should fix that * Remove activationDelay option It complicated the code a lot, and is rather arbitrary. What if your activation script takes more than activationDelay seconds? Instead, one should use systemd dependencies to make sure some action happens before setting the certificate live. e.g. If you want to wait until your cert is published in DNS DANE / TLSA, you could create a unit that blocks until it appears in DNS: ``` RequiredBy=acme-${cert}.service After=acme-${cert}.service ExecStart=publish-wait-for-dns-script ```
2019-08-29 14:32:59 +00:00
locations."/.well-known/acme-challenge".root = "/var/lib/acme/acme-challenges";
};
};
2020-02-09 11:34:17 +00:00
webserver = { nodes, config, pkgs, lib, ... }: {
imports = [ commonConfig ];
networking.firewall.allowedTCPPorts = [ 80 443 ];
networking.nameservers = lib.mkForce [
nodes.dnsserver.config.networking.primaryIPAddress
];
Fix letsencrypt (#60219) * nixos/acme: Fix ordering of cert requests When subsequent certificates would be added, they would not wake up nginx correctly due to target units only being triggered once. We now added more fine-grained systemd dependencies to make sure nginx always is aware of new certificates and doesn't restart too early resulting in a crash. Furthermore, the acme module has been refactored. Mostly to get rid of the deprecated PermissionStartOnly systemd options which were deprecated. Below is a summary of changes made. * Use SERVICE_RESULT to determine status This was added in systemd v232. we don't have to keep track of the EXITCODE ourselves anymore. * Add regression test for requesting mutliple domains * Deprecate 'directory' option We now use systemd's StateDirectory option to manage create and permissions of the acme state directory. * The webroot is created using a systemd.tmpfiles.rules rule instead of the preStart script. * Depend on certs directly By getting rid of the target units, we make sure ordering is correct in the case that you add new certs after already having deployed some. Reason it broke before: acme-certificates.target would be in active state, and if you then add a new cert, it would still be active and hence nginx would restart without even requesting a new cert. Not good! We make the dependencies more fine-grained now. this should fix that * Remove activationDelay option It complicated the code a lot, and is rather arbitrary. What if your activation script takes more than activationDelay seconds? Instead, one should use systemd dependencies to make sure some action happens before setting the certificate live. e.g. If you want to wait until your cert is published in DNS DANE / TLSA, you could create a unit that blocks until it appears in DNS: ``` RequiredBy=acme-${cert}.service After=acme-${cert}.service ExecStart=publish-wait-for-dns-script ```
2019-08-29 14:32:59 +00:00
# A target remains active. Use this to probe the fact that
# a service fired eventhough it is not RemainAfterExit
systemd.targets."acme-finished-a.example.test" = {
after = [ "acme-a.example.test.service" ];
wantedBy = [ "acme-a.example.test.service" ];
};
Fix letsencrypt (#60219) * nixos/acme: Fix ordering of cert requests When subsequent certificates would be added, they would not wake up nginx correctly due to target units only being triggered once. We now added more fine-grained systemd dependencies to make sure nginx always is aware of new certificates and doesn't restart too early resulting in a crash. Furthermore, the acme module has been refactored. Mostly to get rid of the deprecated PermissionStartOnly systemd options which were deprecated. Below is a summary of changes made. * Use SERVICE_RESULT to determine status This was added in systemd v232. we don't have to keep track of the EXITCODE ourselves anymore. * Add regression test for requesting mutliple domains * Deprecate 'directory' option We now use systemd's StateDirectory option to manage create and permissions of the acme state directory. * The webroot is created using a systemd.tmpfiles.rules rule instead of the preStart script. * Depend on certs directly By getting rid of the target units, we make sure ordering is correct in the case that you add new certs after already having deployed some. Reason it broke before: acme-certificates.target would be in active state, and if you then add a new cert, it would still be active and hence nginx would restart without even requesting a new cert. Not good! We make the dependencies more fine-grained now. this should fix that * Remove activationDelay option It complicated the code a lot, and is rather arbitrary. What if your activation script takes more than activationDelay seconds? Instead, one should use systemd dependencies to make sure some action happens before setting the certificate live. e.g. If you want to wait until your cert is published in DNS DANE / TLSA, you could create a unit that blocks until it appears in DNS: ``` RequiredBy=acme-${cert}.service After=acme-${cert}.service ExecStart=publish-wait-for-dns-script ```
2019-08-29 14:32:59 +00:00
services.nginx.enable = true;
Fix letsencrypt (#60219) * nixos/acme: Fix ordering of cert requests When subsequent certificates would be added, they would not wake up nginx correctly due to target units only being triggered once. We now added more fine-grained systemd dependencies to make sure nginx always is aware of new certificates and doesn't restart too early resulting in a crash. Furthermore, the acme module has been refactored. Mostly to get rid of the deprecated PermissionStartOnly systemd options which were deprecated. Below is a summary of changes made. * Use SERVICE_RESULT to determine status This was added in systemd v232. we don't have to keep track of the EXITCODE ourselves anymore. * Add regression test for requesting mutliple domains * Deprecate 'directory' option We now use systemd's StateDirectory option to manage create and permissions of the acme state directory. * The webroot is created using a systemd.tmpfiles.rules rule instead of the preStart script. * Depend on certs directly By getting rid of the target units, we make sure ordering is correct in the case that you add new certs after already having deployed some. Reason it broke before: acme-certificates.target would be in active state, and if you then add a new cert, it would still be active and hence nginx would restart without even requesting a new cert. Not good! We make the dependencies more fine-grained now. this should fix that * Remove activationDelay option It complicated the code a lot, and is rather arbitrary. What if your activation script takes more than activationDelay seconds? Instead, one should use systemd dependencies to make sure some action happens before setting the certificate live. e.g. If you want to wait until your cert is published in DNS DANE / TLSA, you could create a unit that blocks until it appears in DNS: ``` RequiredBy=acme-${cert}.service After=acme-${cert}.service ExecStart=publish-wait-for-dns-script ```
2019-08-29 14:32:59 +00:00
services.nginx.virtualHosts."a.example.test" = {
enableACME = true;
forceSSL = true;
locations."/".root = pkgs.runCommand "docroot" {} ''
mkdir -p "$out"
echo hello world > "$out/index.html"
'';
};
Fix letsencrypt (#60219) * nixos/acme: Fix ordering of cert requests When subsequent certificates would be added, they would not wake up nginx correctly due to target units only being triggered once. We now added more fine-grained systemd dependencies to make sure nginx always is aware of new certificates and doesn't restart too early resulting in a crash. Furthermore, the acme module has been refactored. Mostly to get rid of the deprecated PermissionStartOnly systemd options which were deprecated. Below is a summary of changes made. * Use SERVICE_RESULT to determine status This was added in systemd v232. we don't have to keep track of the EXITCODE ourselves anymore. * Add regression test for requesting mutliple domains * Deprecate 'directory' option We now use systemd's StateDirectory option to manage create and permissions of the acme state directory. * The webroot is created using a systemd.tmpfiles.rules rule instead of the preStart script. * Depend on certs directly By getting rid of the target units, we make sure ordering is correct in the case that you add new certs after already having deployed some. Reason it broke before: acme-certificates.target would be in active state, and if you then add a new cert, it would still be active and hence nginx would restart without even requesting a new cert. Not good! We make the dependencies more fine-grained now. this should fix that * Remove activationDelay option It complicated the code a lot, and is rather arbitrary. What if your activation script takes more than activationDelay seconds? Instead, one should use systemd dependencies to make sure some action happens before setting the certificate live. e.g. If you want to wait until your cert is published in DNS DANE / TLSA, you could create a unit that blocks until it appears in DNS: ``` RequiredBy=acme-${cert}.service After=acme-${cert}.service ExecStart=publish-wait-for-dns-script ```
2019-08-29 14:32:59 +00:00
security.acme.server = "https://acme.test/dir";
specialisation.second-cert.configuration = {pkgs, ...}: {
systemd.targets."acme-finished-b.example.test" = {
after = [ "acme-b.example.test.service" ];
wantedBy = [ "acme-b.example.test.service" ];
};
services.nginx.virtualHosts."b.example.test" = {
enableACME = true;
forceSSL = true;
locations."/".root = pkgs.runCommand "docroot" {} ''
mkdir -p "$out"
echo hello world > "$out/index.html"
'';
};
};
specialisation.dns-01.configuration = {pkgs, config, nodes, lib, ...}: {
security.acme.certs."example.test" = {
domain = "*.example.test";
dnsProvider = "exec";
dnsPropagationCheck = false;
credentialsFile = with pkgs; writeText "wildcard.env" ''
EXEC_PATH=${dnsScript { inherit writeScript bash curl; dnsAddress = nodes.dnsserver.config.networking.primaryIPAddress; }}
'';
user = config.services.nginx.user;
group = config.services.nginx.group;
};
systemd.targets."acme-finished-example.test" = {
after = [ "acme-example.test.service" ];
wantedBy = [ "acme-example.test.service" ];
};
systemd.services."acme-example.test" = {
before = [ "nginx.service" ];
wantedBy = [ "nginx.service" ];
};
services.nginx.virtualHosts."c.example.test" = {
forceSSL = true;
sslCertificate = config.security.acme.certs."example.test".directory + "/cert.pem";
sslTrustedCertificate = config.security.acme.certs."example.test".directory + "/full.pem";
sslCertificateKey = config.security.acme.certs."example.test".directory + "/key.pem";
locations."/".root = pkgs.runCommand "docroot" {} ''
mkdir -p "$out"
echo hello world > "$out/index.html"
'';
};
};
# When nginx depends on a service that is slow to start up, requesting used to fail
# certificates fail. Reproducer for https://github.com/NixOS/nixpkgs/issues/81842
specialisation.slow-startup.configuration = { pkgs, config, nodes, lib, ...}: {
systemd.services.my-slow-service = {
wantedBy = [ "multi-user.target" "nginx.service" ];
before = [ "nginx.service" ];
preStart = "sleep 5";
script = "${pkgs.python3}/bin/python -m http.server";
};
systemd.targets."acme-finished-d.example.com" = {
after = [ "acme-d.example.com.service" ];
wantedBy = [ "acme-d.example.com.service" ];
};
services.nginx.virtualHosts."d.example.com" = {
forceSSL = true;
enableACME = true;
locations."/".proxyPass = "http://localhost:8000";
};
};
};
client = {nodes, lib, ...}: {
imports = [ commonConfig ];
networking.nameservers = lib.mkForce [
nodes.dnsserver.config.networking.primaryIPAddress
];
};
};
testScript = {nodes, ...}:
Fix letsencrypt (#60219) * nixos/acme: Fix ordering of cert requests When subsequent certificates would be added, they would not wake up nginx correctly due to target units only being triggered once. We now added more fine-grained systemd dependencies to make sure nginx always is aware of new certificates and doesn't restart too early resulting in a crash. Furthermore, the acme module has been refactored. Mostly to get rid of the deprecated PermissionStartOnly systemd options which were deprecated. Below is a summary of changes made. * Use SERVICE_RESULT to determine status This was added in systemd v232. we don't have to keep track of the EXITCODE ourselves anymore. * Add regression test for requesting mutliple domains * Deprecate 'directory' option We now use systemd's StateDirectory option to manage create and permissions of the acme state directory. * The webroot is created using a systemd.tmpfiles.rules rule instead of the preStart script. * Depend on certs directly By getting rid of the target units, we make sure ordering is correct in the case that you add new certs after already having deployed some. Reason it broke before: acme-certificates.target would be in active state, and if you then add a new cert, it would still be active and hence nginx would restart without even requesting a new cert. Not good! We make the dependencies more fine-grained now. this should fix that * Remove activationDelay option It complicated the code a lot, and is rather arbitrary. What if your activation script takes more than activationDelay seconds? Instead, one should use systemd dependencies to make sure some action happens before setting the certificate live. e.g. If you want to wait until your cert is published in DNS DANE / TLSA, you could create a unit that blocks until it appears in DNS: ``` RequiredBy=acme-${cert}.service After=acme-${cert}.service ExecStart=publish-wait-for-dns-script ```
2019-08-29 14:32:59 +00:00
let
newServerSystem = nodes.webserver.config.system.build.toplevel;
Fix letsencrypt (#60219) * nixos/acme: Fix ordering of cert requests When subsequent certificates would be added, they would not wake up nginx correctly due to target units only being triggered once. We now added more fine-grained systemd dependencies to make sure nginx always is aware of new certificates and doesn't restart too early resulting in a crash. Furthermore, the acme module has been refactored. Mostly to get rid of the deprecated PermissionStartOnly systemd options which were deprecated. Below is a summary of changes made. * Use SERVICE_RESULT to determine status This was added in systemd v232. we don't have to keep track of the EXITCODE ourselves anymore. * Add regression test for requesting mutliple domains * Deprecate 'directory' option We now use systemd's StateDirectory option to manage create and permissions of the acme state directory. * The webroot is created using a systemd.tmpfiles.rules rule instead of the preStart script. * Depend on certs directly By getting rid of the target units, we make sure ordering is correct in the case that you add new certs after already having deployed some. Reason it broke before: acme-certificates.target would be in active state, and if you then add a new cert, it would still be active and hence nginx would restart without even requesting a new cert. Not good! We make the dependencies more fine-grained now. this should fix that * Remove activationDelay option It complicated the code a lot, and is rather arbitrary. What if your activation script takes more than activationDelay seconds? Instead, one should use systemd dependencies to make sure some action happens before setting the certificate live. e.g. If you want to wait until your cert is published in DNS DANE / TLSA, you could create a unit that blocks until it appears in DNS: ``` RequiredBy=acme-${cert}.service After=acme-${cert}.service ExecStart=publish-wait-for-dns-script ```
2019-08-29 14:32:59 +00:00
switchToNewServer = "${newServerSystem}/bin/switch-to-configuration test";
in
2019-10-27 12:53:55 +00:00
# Note, wait_for_unit does not work for oneshot services that do not have RemainAfterExit=true,
Fix letsencrypt (#60219) * nixos/acme: Fix ordering of cert requests When subsequent certificates would be added, they would not wake up nginx correctly due to target units only being triggered once. We now added more fine-grained systemd dependencies to make sure nginx always is aware of new certificates and doesn't restart too early resulting in a crash. Furthermore, the acme module has been refactored. Mostly to get rid of the deprecated PermissionStartOnly systemd options which were deprecated. Below is a summary of changes made. * Use SERVICE_RESULT to determine status This was added in systemd v232. we don't have to keep track of the EXITCODE ourselves anymore. * Add regression test for requesting mutliple domains * Deprecate 'directory' option We now use systemd's StateDirectory option to manage create and permissions of the acme state directory. * The webroot is created using a systemd.tmpfiles.rules rule instead of the preStart script. * Depend on certs directly By getting rid of the target units, we make sure ordering is correct in the case that you add new certs after already having deployed some. Reason it broke before: acme-certificates.target would be in active state, and if you then add a new cert, it would still be active and hence nginx would restart without even requesting a new cert. Not good! We make the dependencies more fine-grained now. this should fix that * Remove activationDelay option It complicated the code a lot, and is rather arbitrary. What if your activation script takes more than activationDelay seconds? Instead, one should use systemd dependencies to make sure some action happens before setting the certificate live. e.g. If you want to wait until your cert is published in DNS DANE / TLSA, you could create a unit that blocks until it appears in DNS: ``` RequiredBy=acme-${cert}.service After=acme-${cert}.service ExecStart=publish-wait-for-dns-script ```
2019-08-29 14:32:59 +00:00
# this is because a oneshot goes from inactive => activating => inactive, and never
# reaches the active state. To work around this, we create some mock target units which
# get pulled in by the oneshot units. The target units linger after activation, and hence we
# can use them to probe that a oneshot fired. It is a bit ugly, but it is the best we can do
''
2019-10-27 12:53:55 +00:00
client.start()
dnsserver.start()
acme.wait_for_unit("default.target")
dnsserver.wait_for_unit("pebble-challtestsrv.service")
client.succeed(
'curl --data \'{"host": "acme.test", "addresses": ["${nodes.acme.config.networking.primaryIPAddress}"]}\' http://${nodes.dnsserver.config.networking.primaryIPAddress}:8055/add-a'
)
client.succeed(
'curl --data \'{"host": "standalone.test", "addresses": ["${nodes.acmeStandalone.config.networking.primaryIPAddress}"]}\' http://${nodes.dnsserver.config.networking.primaryIPAddress}:8055/add-a'
)
acme.start()
2019-10-27 12:53:55 +00:00
acmeStandalone.start()
acme.wait_for_unit("default.target")
acme.wait_for_unit("pebble.service")
2019-10-27 12:53:55 +00:00
with subtest("can request certificate with HTTPS-01 challenge"):
acmeStandalone.wait_for_unit("default.target")
acmeStandalone.succeed("systemctl start acme-standalone.test.service")
acmeStandalone.wait_for_unit("acme-finished-standalone.test.target")
2019-10-27 12:53:55 +00:00
client.wait_for_unit("default.target")
client.succeed("curl https://acme.test:15000/roots/0 > /tmp/ca.crt")
client.succeed("curl https://acme.test:15000/intermediate-keys/0 >> /tmp/ca.crt")
2019-10-27 12:53:55 +00:00
with subtest("Can request certificate for nginx service"):
webserver.wait_for_unit("acme-finished-a.example.test.target")
2019-10-27 12:53:55 +00:00
client.succeed(
"curl --cacert /tmp/ca.crt https://a.example.test/ | grep -qF 'hello world'"
2019-10-27 12:53:55 +00:00
)
with subtest("Can add another certificate for nginx service"):
webserver.succeed(
"/run/current-system/specialisation/second-cert/bin/switch-to-configuration test"
2019-10-27 12:53:55 +00:00
)
webserver.wait_for_unit("acme-finished-b.example.test.target")
2019-10-27 12:53:55 +00:00
client.succeed(
"curl --cacert /tmp/ca.crt https://b.example.test/ | grep -qF 'hello world'"
2019-10-27 12:53:55 +00:00
)
with subtest("Can request wildcard certificates using DNS-01 challenge"):
webserver.succeed(
"${switchToNewServer}"
)
webserver.succeed(
"/run/current-system/specialisation/dns-01/bin/switch-to-configuration test"
)
webserver.wait_for_unit("acme-finished-example.test.target")
client.succeed(
"curl --cacert /tmp/ca.crt https://c.example.test/ | grep -qF 'hello world'"
)
with subtest("Can request certificate of nginx when startup is delayed"):
webserver.succeed(
"${switchToNewServer}"
)
webserver.succeed(
"/run/current-system/specialisation/slow-startup/bin/switch-to-configuration test"
)
webserver.wait_for_unit("acme-finished-d.example.com.target")
client.succeed("curl --cacert /tmp/ca.crt https://d.example.com/")
Fix letsencrypt (#60219) * nixos/acme: Fix ordering of cert requests When subsequent certificates would be added, they would not wake up nginx correctly due to target units only being triggered once. We now added more fine-grained systemd dependencies to make sure nginx always is aware of new certificates and doesn't restart too early resulting in a crash. Furthermore, the acme module has been refactored. Mostly to get rid of the deprecated PermissionStartOnly systemd options which were deprecated. Below is a summary of changes made. * Use SERVICE_RESULT to determine status This was added in systemd v232. we don't have to keep track of the EXITCODE ourselves anymore. * Add regression test for requesting mutliple domains * Deprecate 'directory' option We now use systemd's StateDirectory option to manage create and permissions of the acme state directory. * The webroot is created using a systemd.tmpfiles.rules rule instead of the preStart script. * Depend on certs directly By getting rid of the target units, we make sure ordering is correct in the case that you add new certs after already having deployed some. Reason it broke before: acme-certificates.target would be in active state, and if you then add a new cert, it would still be active and hence nginx would restart without even requesting a new cert. Not good! We make the dependencies more fine-grained now. this should fix that * Remove activationDelay option It complicated the code a lot, and is rather arbitrary. What if your activation script takes more than activationDelay seconds? Instead, one should use systemd dependencies to make sure some action happens before setting the certificate live. e.g. If you want to wait until your cert is published in DNS DANE / TLSA, you could create a unit that blocks until it appears in DNS: ``` RequiredBy=acme-${cert}.service After=acme-${cert}.service ExecStart=publish-wait-for-dns-script ```
2019-08-29 14:32:59 +00:00
'';
})