Each locks <home>/run/<role>.lock in main (gatewayd after its secret, so nothing is made before the secret is read). brokerd's old second-instance test in serve.rs expected the audit lock's message, so serve.rs has a new version; task 01 now copies serve.rs-task01. Checks: given tests fail at ref 12 as the task says (2 each in brokerd, memoryd, gatewayd; serve.rs passes); the task's quoted edits applied to ref 12 reproduce ref 13's tree exactly; replay gate ok with 1118 tests (1113 - 1 + 6), /tmp unchanged; 14 mutations, all caught; the six instance test files stressed 25 runs under 16 busy loops and 20 under 32, no failure. Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
219 lines
6.8 KiB
Rust
219 lines
6.8 KiB
Rust
//! One `gatewayd` per home (M5b spec, section 9a): a second `gatewayd serve` against the same home
|
|
//! is refused with the lock's line, and the first keeps serving `gateway.sock`; once the first is
|
|
//! killed, a new one removes the stale socket and serves. Each runs against the fake Mattermost.
|
|
//! Every wait has a deadline, and every process is killed and every path removed by a guard (tip
|
|
//! T44). Do not edit.
|
|
|
|
#[path = "support/brokerd_client.rs"]
|
|
mod brokerd_client;
|
|
#[path = "support/fake_mm.rs"]
|
|
mod fake_mm;
|
|
#[path = "support/tmp.rs"]
|
|
mod tmp;
|
|
|
|
use std::os::unix::fs::PermissionsExt;
|
|
use std::path::{Path, PathBuf};
|
|
use std::process::{Child, Command, Stdio};
|
|
use std::time::{Duration, Instant};
|
|
|
|
use brokerd_client::{ask, pending};
|
|
use fake_mm::{FakeMm, KYLE};
|
|
use tmp::TempDir;
|
|
|
|
const POINTER: &str = "see docs/runbook.md#already-running";
|
|
const TOKEN: &str = "tok-instance-test";
|
|
const WAIT: Duration = Duration::from_secs(10);
|
|
|
|
struct Home {
|
|
dir: TempDir,
|
|
config: PathBuf,
|
|
}
|
|
|
|
impl Home {
|
|
fn new(tag: &str, mm: &FakeMm) -> Home {
|
|
let dir = TempDir::new(tag);
|
|
let text = format!(
|
|
"[mattermost]\nurl = \"{}\"\n[secrets.mattermost_token]\nenv = \"GW_TEST_TOKEN\"\n\
|
|
[allow]\nusers = [\"{KYLE}\"]\n[paths]\nhome = \"{}\"\n",
|
|
mm.url(),
|
|
dir.path().display()
|
|
);
|
|
let config = dir.write("gatewayd.toml", &text);
|
|
Home { dir, config }
|
|
}
|
|
|
|
fn path(&self, relative: &str) -> PathBuf {
|
|
self.dir.path().join(relative)
|
|
}
|
|
|
|
fn socket(&self) -> PathBuf {
|
|
self.path("run/gateway/gateway.sock")
|
|
}
|
|
|
|
fn command(&self) -> Command {
|
|
let mut command = Command::new(env!("CARGO_BIN_EXE_gatewayd"));
|
|
command
|
|
.args(["serve", "--config"])
|
|
.arg(&self.config)
|
|
.env_remove("CREDENTIALS_DIRECTORY")
|
|
.env("GW_TEST_TOKEN", TOKEN)
|
|
.stdin(Stdio::null())
|
|
.stdout(Stdio::null());
|
|
command
|
|
}
|
|
|
|
/// `gatewayd serve` with its stderr in `<home>/<name>.err`, for the messages of a failing test.
|
|
fn start(&self, name: &str) -> Daemon {
|
|
let err = std::fs::File::create(self.path(&format!("{name}.err"))).unwrap();
|
|
Daemon(self.command().stderr(Stdio::from(err)).spawn().unwrap())
|
|
}
|
|
|
|
fn stderr_of(&self, name: &str) -> String {
|
|
std::fs::read_to_string(self.path(&format!("{name}.err"))).unwrap_or_default()
|
|
}
|
|
|
|
/// Runs `gatewayd serve` to its end, at most 10 s (then it is killed), and gives its exit code
|
|
/// (`None` if it had to be killed) and its stderr.
|
|
fn run_to_end(&self) -> (Option<i32>, String) {
|
|
let mut child = Daemon(self.command().stderr(Stdio::piped()).spawn().unwrap());
|
|
let deadline = Instant::now() + WAIT;
|
|
let mut code = None;
|
|
while Instant::now() < deadline {
|
|
if let Ok(Some(status)) = child.0.try_wait() {
|
|
code = status.code();
|
|
break;
|
|
}
|
|
std::thread::sleep(Duration::from_millis(20));
|
|
}
|
|
let _ = child.0.kill();
|
|
let _ = child.0.wait();
|
|
let mut stderr = String::new();
|
|
if let Some(mut pipe) = child.0.stderr.take() {
|
|
let _ = std::io::Read::read_to_string(&mut pipe, &mut stderr);
|
|
}
|
|
(code, stderr)
|
|
}
|
|
}
|
|
|
|
/// A running `gatewayd serve`, killed and waited for when dropped. Declared after the `Home` and
|
|
/// the `FakeMm`, so it drops first.
|
|
struct Daemon(Child);
|
|
|
|
impl Drop for Daemon {
|
|
fn drop(&mut self) {
|
|
let _ = self.0.kill();
|
|
let _ = self.0.wait();
|
|
}
|
|
}
|
|
|
|
impl Daemon {
|
|
fn running(&mut self) -> bool {
|
|
matches!(self.0.try_wait(), Ok(None))
|
|
}
|
|
}
|
|
|
|
/// True when `gatewayd` answers on `gateway.sock` as it does for `brokerd`: approval `approval`,
|
|
/// of a `bxctl chat` session, is posted to the first approver and the posts come back. The
|
|
/// connection is then closed, which ends the approval.
|
|
fn answers(socket: &Path, approval: u64) -> bool {
|
|
match ask(socket, pending(approval, "chat-1789700000-1", "{}")) {
|
|
Ok(asking) => {
|
|
let answered = !asking.posts.is_empty();
|
|
asking.close();
|
|
answered
|
|
}
|
|
Err(_) => false,
|
|
}
|
|
}
|
|
|
|
/// Waits up to 10 s for `gatewayd` to answer on `socket`.
|
|
fn wait_until_it_answers(socket: &Path, approval: u64) -> bool {
|
|
let deadline = Instant::now() + WAIT;
|
|
while Instant::now() < deadline {
|
|
if answers(socket, approval) {
|
|
return true;
|
|
}
|
|
std::thread::sleep(Duration::from_millis(50));
|
|
}
|
|
false
|
|
}
|
|
|
|
fn inode(path: &Path) -> u64 {
|
|
use std::os::unix::fs::MetadataExt;
|
|
std::fs::symlink_metadata(path)
|
|
.map(|m| m.ino())
|
|
.unwrap_or(0)
|
|
}
|
|
|
|
#[test]
|
|
fn a_second_gatewayd_on_the_same_home_is_refused_and_the_first_keeps_serving() {
|
|
let mm = FakeMm::start();
|
|
let home = Home::new("inst-twice", &mm);
|
|
let mut first = home.start("first");
|
|
let _ws = mm.next_ws(WAIT);
|
|
assert!(
|
|
wait_until_it_answers(&home.socket(), 1),
|
|
"the first gatewayd never answered; stderr: {}",
|
|
home.stderr_of("first")
|
|
);
|
|
let before = inode(&home.socket());
|
|
|
|
let (code, stderr) = home.run_to_end();
|
|
let line = format!(
|
|
"gatewayd: another gatewayd is running (it holds {}); {POINTER}",
|
|
home.path("run/gatewayd.lock").display()
|
|
);
|
|
assert_eq!(code, Some(1), "the second gatewayd: {stderr}");
|
|
assert_eq!(
|
|
stderr,
|
|
format!("{line}\n"),
|
|
"the lock's line, and nothing else"
|
|
);
|
|
|
|
assert!(first.running(), "the first gatewayd is still running");
|
|
assert_eq!(
|
|
inode(&home.socket()),
|
|
before,
|
|
"the first's socket is still there"
|
|
);
|
|
assert!(
|
|
answers(&home.socket(), 2),
|
|
"the first gatewayd still answers; its stderr: {}",
|
|
home.stderr_of("first")
|
|
);
|
|
}
|
|
|
|
#[test]
|
|
fn after_the_first_is_killed_a_new_gatewayd_removes_the_stale_socket_and_serves() {
|
|
let mm = FakeMm::start();
|
|
let home = Home::new("inst-restart", &mm);
|
|
let mut first = home.start("first");
|
|
let _ws = mm.next_ws(WAIT);
|
|
assert!(
|
|
wait_until_it_answers(&home.socket(), 1),
|
|
"the first gatewayd never answered; stderr: {}",
|
|
home.stderr_of("first")
|
|
);
|
|
let _ = first.0.kill();
|
|
let _ = first.0.wait();
|
|
assert!(
|
|
home.socket().exists(),
|
|
"a killed gatewayd leaves its socket"
|
|
);
|
|
assert!(
|
|
!answers(&home.socket(), 2),
|
|
"nobody serves the stale socket"
|
|
);
|
|
|
|
let _next = home.start("next");
|
|
let _ws = mm.next_ws(WAIT);
|
|
assert!(
|
|
wait_until_it_answers(&home.socket(), 3),
|
|
"the new gatewayd never answered; stderr: {}",
|
|
home.stderr_of("next")
|
|
);
|
|
let lock = home.path("run/gatewayd.lock");
|
|
let mode = std::fs::metadata(&lock).map(|m| m.permissions().mode() & 0o777);
|
|
assert_eq!(mode.ok(), Some(0o600), "the lock file, mode 0600");
|
|
}
|