facebook/flow · error

failed to write status

Error message

failed to write status

What it means

Raised by print_status when the non-spinner branch, writeln!(stderr, "{}", message), fails to write the plain status message to stderr. This branch runs whenever progress/spinner output is disabled or stderr is not a terminal, and the expect panics because the library assumes status messages can always be written. Failure indicates stderr is closed, redirected to a broken pipe, or hitting an I/O error.

Source

Thrown at rust_port/crates/flow_commands_connect/src/command_connect.rs:52

    pub emoji: bool,
    pub quiet: bool,
    pub show_progress: bool,
    pub flowconfig_name: &'a str,
    pub rerun_on_mismatch: bool,
}

fn print_status(env: &Env<'_>, message: &str) {
    if env.quiet {
        return;
    }

    let stderr = std::io::stderr();
    let mut stderr = stderr.lock();
    if env.show_progress && stderr.is_terminal() {
        write!(stderr, "{}: {}", message, flow_utils_tty::spinner(false))
            .expect("failed to write spinner status");
    } else {
        writeln!(stderr, "{}", message).expect("failed to write status");
    }
    stderr.flush().expect("failed to flush status");
}

fn arg(name: &str, value: Option<&str>, arr: &mut Vec<String>) {
    if let Some(value) = value {
        arr.push(name.to_string());
        arr.push(value.to_string());
    }
}

fn flag(name: &str, value: bool, arr: &mut Vec<String>) {
    if value {
        arr.push(name.to_string());
    }
}

// Starts up a flow server by literally calling flow start

View on GitHub (pinned to 5c86586199)

Solutions

1. Redirect stderr to a writable target (file or live terminal) and rerun.
2. Free disk space if stderr is a log file on a full filesystem.
3. Stop piping stderr into consumers that terminate before the command completes.
4. Patch the library to tolerate stderr write failures (log-and-continue) if you control the code.

Example fix

// invocation, before
flow check 2>&1 | head -n 5
// after
flow check 2>flow-check.log; head -n 5 flow-check.log

use std::io::Write;
fn stderr_writable() -> bool {
    let mut e = std::io::stderr();
    e.write_all(b"").and_then(|_| e.flush()).is_ok()
}
// call before invoking; skip/redirect if false

fn stderr_alive() -> bool { stderr_writable() }

match std::panic::catch_unwind(|| connect_flow()) {
    Ok(r) => r,
    Err(_) => { let _ = std::fs::write("flow-status-fallback.log", "status write failed"); std::process::exit(1); }
}

When it happens

Trigger: connect_rec or handle_missing_server calls print_status with show_progress off (or stderr not a TTY) while stderr's writer is gone: closed fd 2, piped into a terminated consumer, or disk full when stderr points to a file.

Common situations: CI jobs where stderr is redirected to a log file on a full disk; running under wrappers that close stderr; piping output to `head`/`grep -q` that exits early; backgrounding the process with stderr attached to a dead terminal.

Related errors

• failed to write spinner status
• failed to flush status
• failed to clear spinner line
• failed to write spinner status
• failed to flush spinner status

AI-assisted analysis of facebook/flow@5c86586199 (2026-09-08). Data as JSON: /api/errors/f81855419a835301. Report an issue: GitHub.