Skip to content

Fix serve-agent reinstall race in schedule:weekly #7

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
howieyoung merged 1 commit into from
Aug 8, 2026
Merged

Fix serve-agent reinstall race in schedule:weekly #7

Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
32 changes: 26 additions & 6 deletions scripts/install-weekly.mjs
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -136,16 +136,36 @@ const uid = process.getuid();
const laDir = path.join(home, "Library", "LaunchAgents");
fs.mkdirSync(laDir, { recursive: true });

// Synchronous sleep (no sleep binary, no async in this install flow)
const sleepMs = (ms) => Atomics.wait(new Int32Array(new SharedArrayBuffer(4)), 0, 0, ms);

// Bootout, then wait until launchd has actually released the label. `bootout` returns before teardown
// finishes, and bootstrapping into a domain that still holds the (KeepAlive) label races and fails with
// "Bootstrap failed: 5: Input/output error", leaving the resident receiver DOWN. Poll until it's gone.
function bootoutAndWait(agentLabel) {
spawnSync("launchctl", ["bootout", `gui/${uid}/${agentLabel}`], { stdio: "ignore" }); // may already be absent; that's fine
for (let i = 0; i < 30; i++) {
const present = spawnSync("launchctl", ["print", `gui/${uid}/${agentLabel}`], { stdio: "ignore" });
if (present.status !== 0) return; // no longer registered → safe to bootstrap
sleepMs(100);
}
}

function installAgent(agentLabel, xml) {
const plistPath = path.join(laDir, `${agentLabel}.plist`);
fs.writeFileSync(plistPath, xml);
spawnSync("launchctl", ["bootout", `gui/${uid}/${agentLabel}`], { stdio: "ignore" }); // bootout the old version first; failure is fine
const boot = spawnSync("launchctl", ["bootstrap", `gui/${uid}`, plistPath], { encoding: "utf8" });
if (boot.status !== 0) {
console.error(`launchctl bootstrap ${agentLabel} failed: ${boot.stderr || boot.stdout}`);
process.exit(1);
bootoutAndWait(agentLabel);
// Even after the label clears, launchd can briefly return a transient EIO; retry a few times before giving up.
let boot;
for (let attempt = 1; attempt <= 5; attempt++) {
boot = spawnSync("launchctl", ["bootstrap", `gui/${uid}`, plistPath], { encoding: "utf8" });
if (boot.status === 0) return plistPath;
if (!/Input\/output error|Bootstrap failed|already/i.test(boot.stderr || boot.stdout || "")) break; // non-transient → fail now
sleepMs(300);
bootoutAndWait(agentLabel); // clear any half-registered state, then try again
}
return plistPath;
console.error(`launchctl bootstrap ${agentLabel} failed: ${boot.stderr || boot.stdout}`);
process.exit(1);
}

// Agents run through a wrapper that resolves node at runtime (no baked node path), executing
Expand Down
Loading