Files
lix/tests/nixos/remote-builds-ssh-ng.nix
T
Raito Bezarius 33eaaf02fd libstore/ssh: remove echo started check
The pre-flight `echo started` check over SSH was originally added in
577ebeaefb. As it is usual with these old
commits, understanding why is there a need for something is difficult.

The closest thing would be
> Fix a race starting the SSH master. We now wait synchronously for
> the SSH master to finish starting. This prevents the SSH clients
> from starting their own connections.

But, we removed SSH connection sharing, so this does not apply anymore.

Nonetheless, we believed this check was meant as a way to catch obvious
misconfigurations or SSH failures early, before handing off to
`nix-store`. However, this approach was not fruitful: it assumes the
remote has a `bash`-compatible shell, `echo` behaves in a standard way,
and no `ForceCommand` interferes—all of which are unreliable assumptions
in practice.

While the intent was to provide slightly better diagnostics (e.g. in
case of SSH hanging or returning an interactive shell), in practice it
does not meaningfully catch or improve real failure cases. The
underlying protocol or engine can and should handle those errors more
robustly anyway.

In contrast, this check *does* break several legitimate workflows,
including:

* remote builders using `ForceCommand` wrappers (e.g.
`nix-remote-build`-style setups), see
<https://discourse.nixos.org/t/wrapper-to-restrict-builder-access-through-ssh-worth-upstreaming/25834/15>,

* SSHing into minimal environments lacking `bash` (e.g. initrd,
busybox-based systems),

* configurations that don’t default to POSIX-like shells, e.g., nushell
enthusiasts.

As such, we’re removing this code. Protocol mismatch errors and SSH
failures can be rethought and handled more structurally elsewhere in the
engine.

Change-Id: I187f6881375d42ef83987a13a350c97964bbdb30
Signed-off-by: Raito Bezarius <raito@lix.systems>
(cherry picked from commit 0dd8bf6c1c)
2025-05-18 19:51:41 +00:00

132 lines
4.4 KiB
Nix

test@{ config, lib, hostPkgs, ... }:
let
pkgs = config.nodes.client.nixpkgs.pkgs;
# Trivial Nix expression to build remotely.
expr = config: nr: pkgs.writeText "expr.nix"
''
let utils = builtins.storePath ${config.system.build.extraUtils}; in
derivation {
name = "hello-${toString nr}";
system = "i686-linux";
PATH = "''${utils}/bin";
builder = "''${utils}/bin/sh";
args = [ "-c" "${
lib.concatStringsSep "; " [
''if [[ -n $NIX_LOG_FD ]]''
''then echo '@nix {\"action\":\"setPhase\",\"phase\":\"buildPhase\"}' >&''$NIX_LOG_FD''
"fi"
"echo Hello"
"mkdir $out"
"cat /proc/sys/kernel/hostname > $out/host"
]
}" ];
outputs = [ "out" ];
}
'';
in
{
options = {
builders.config = lib.mkOption {
type = lib.types.deferredModule;
description = ''
Configuration to add to the builder nodes.
'';
default = { };
};
};
config = {
name = lib.mkDefault "remote-builds-ssh-ng";
nodes =
{ builder =
{ config, pkgs, ... }:
{
imports = [ test.config.builders.config ];
services.openssh.enable = true;
virtualisation.writableStore = true;
nix.settings.sandbox = true;
nix.settings.substituters = lib.mkForce [ ];
};
client =
{ config, lib, pkgs, ... }:
{ nix.settings.max-jobs = 0; # force remote building
nix.distributedBuilds = true;
nix.buildMachines =
[ { hostName = "builder";
sshUser = "root";
sshKey = "/root/.ssh/id_ed25519";
system = "i686-linux";
maxJobs = 1;
protocol = "ssh-ng";
}
];
virtualisation.writableStore = true;
virtualisation.additionalPaths = [ config.system.build.extraUtils ];
nix.settings.substituters = lib.mkForce [ ];
programs.ssh.extraConfig = "ConnectTimeout 30";
};
};
testScript = { nodes }: ''
# fmt: off
import subprocess
start_all()
builder.succeed("systemctl start network-online.target")
client.succeed("systemctl start network-online.target")
builder.wait_for_unit("network-online.target")
client.wait_for_unit("network-online.target")
# Create an SSH key on the client.
subprocess.run([
"${hostPkgs.openssh}/bin/ssh-keygen", "-t", "ed25519", "-f", "key", "-N", ""
], capture_output=True, check=True)
client.succeed("mkdir -p -m 700 /root/.ssh")
client.copy_from_host("key", "/root/.ssh/id_ed25519")
client.succeed("chmod 600 /root/.ssh/id_ed25519")
# Install the SSH key on the builder.
builder.succeed("mkdir -p -m 700 /root/.ssh")
builder.copy_from_host("key.pub", "/root/.ssh/authorized_keys")
builder.wait_for_unit("sshd.service")
out = client.fail("nix-build ${expr nodes.client 1} 2>&1")
assert "Host key verification failed." in out, f"No host verification error:\n{out}"
assert "'ssh-ng://root@builder'" in out, f"No details about which host:\n{out}"
client.succeed(f"ssh -o StrictHostKeyChecking=no {builder.name} 'echo hello world' >&2")
# Perform a build
out = client.succeed("nix-build ${expr nodes.client 1} 2> build-output")
# Verify that the build was done on the builder
builder.succeed(f"test -e {out.strip()}")
# Print the build log, prefix the log lines to avoid nix intercepting lines starting with @nix
buildOutput = client.succeed("sed -e 's/^/build-output:/' build-output")
print(buildOutput)
# Make sure that we get the expected build output
client.succeed("grep -qF Hello build-output")
# We don't want phase reporting in the build output
client.fail("grep -qF '@nix' build-output")
# Get the log file
client.succeed(f"nix-store --read-log {out.strip()} > log-output")
# Prefix the log lines to avoid nix intercepting lines starting with @nix
logOutput = client.succeed("sed -e 's/^/log-file:/' log-output")
print(logOutput)
# Check that we get phase reporting in the log file
client.succeed("grep -q '@nix {\"action\":\"setPhase\",\"phase\":\"buildPhase\"}' log-output")
'';
};
}