Bash实时读取日志文件并在错误触发时执行命令的问题

Let's break down why your original command isn't working as expected, then fix it properly.

The Problem With Your Original Command #

Your command:

$(tail -n0 -f /Users/user/Desktop/file | grep error | echo 2 > /Users/user/Desktop/error.log) &

has two critical issues:

1. The echo 2 runs unconditionally: Pipes in bash execute each command independently. The echo 2 doesn't wait for or react to the output of grep error — it just runs immediately when the command starts, which is why you always see a 2 written to error.log right away, regardless of actual errors.
2. Unnecessary command substitution ($(...)): Wrapping the whole pipeline in $(...) captures its output (which is nothing here) and tries to run it as a command. This adds no value and just complicates things.

The Fixed Solution #

To only write 2 to error.log when a line containing "error" appears in your target log, use a loop that reacts to matching lines. Here's a reliable version:

tail -n0 -f /Users/user/Desktop/file | grep --line-buffered "error" | while read -r; do
    echo 2 >> /Users/user/Desktop/error.log
done &

Or if you prefer a more concise one-liner:

tail -n0 -f /Users/user/Desktop/file | grep --line-buffered "error" | xargs -I {} echo 2 >> /Users/user/Desktop/error.log &

Key Details Explained #

• --line-buffered: By default, grep buffers output when connected to a pipe, which means it might not send matching lines immediately. This flag forces it to flush output line-by-line, critical for real-time detection.
• while read -r: This loop runs every time grep outputs a matching line, triggering the echo 2 only when an error is detected.
• >> instead of >: Using append (>>) ensures you don't overwrite existing entries in error.log each time an error is found. Use > only if you want to reset the log on each run.

Alternative: Using if with Grep's Exit Code #

If you want to check for errors and write once per detection cycle (instead of once per error line), you could use this approach:

while true; do
    if tail -n0 -f /Users/user/Desktop/file | grep --line-buffered -q "error"; then
        echo 2 >> /Users/user/Desktop/error.log
        # Optional: Break here if you only want to detect the first error
        # break
    fi
done &

The -q flag makes grep quiet (no output) and returns an exit code of 0 only if a match is found, which triggers the echo 2.

内容的提问来源于stack exchange，提问作者madsantos