Heroku部署Jar生成config.txt后，如何让Shell脚本读取该文件？

Got it, let's break down how to make your shell script read that auto-generated config.txt file on Heroku. Here's a step-by-step approach that handles the timing of the Jar's startup and file creation:

1. Adjust Your Shell Script to Wait for config.txt #

The key challenge here is that your Jar needs a moment to generate config.txt before the script can read it. We'll start the Jar in the background, wait for the file to exist, then extract the credentials.

Here's a complete start.sh script example:

#!/bin/bash

# Launch your Jar in the background so the script can continue executing
java -jar your-app.jar &
# Capture the Jar's process ID to manage it later
JAR_PID=$!

# Wait for config.txt to be created (max 30 seconds - adjust as needed)
MAX_WAIT_SECONDS=30
wait_counter=0

while [ ! -f config.txt ]; do
  sleep 1
  wait_counter=$((wait_counter + 1))
  
  # Exit with error if file never shows up
  if [ $wait_counter -ge $MAX_WAIT_SECONDS ]; then
    echo "Error: config.txt wasn't generated within $MAX_WAIT_SECONDS seconds"
    kill $JAR_PID
    exit 1
  fi
done

# Extract username and password from config.txt
# Adjust the grep patterns to match your config.txt's actual format
# Example assumes lines like "username: my-random-user" or "password: my-secret-pass"
USERNAME=$(grep -oP 'username: \K.*' config.txt)
PASSWORD=$(grep -oP 'password: \K.*' config.txt)

# Validate we actually got the credentials
if [ -z "$USERNAME" ] || [ -z "$PASSWORD" ]; then
  echo "Error: Failed to pull username/password from config.txt"
  kill $JAR_PID
  exit 1
fi

# Use the credentials however you need (log them, set env vars, etc.)
echo "Successfully loaded credentials: Username=$USERNAME, Password=***"

# Keep the script running (Heroku kills dynos without a foreground process)
wait $JAR_PID

2. Key Details to Note #

• Background Execution: Starting the Jar with & lets the script proceed to the waiting step instead of blocking indefinitely.
• Wait Logic: The loop ensures we don't try to read the file before it exists. Adjust MAX_WAIT_SECONDS if your Jar takes longer to generate the file.
• Credential Extraction: The grep -oP command uses Perl-compatible regex to capture the value after the label. If your config.txt uses a different format, tweak this part:
  • For username=my-user format: USERNAME=$(grep -oP 'username=\K.*' config.txt)
  • For JSON (you'll need to install jq via your buildpack first): USERNAME=$(jq -r '.username' config.txt)
• Heroku Foreground Requirement: The final wait $JAR_PID keeps the script running as a foreground process, which Heroku requires to keep your dyno alive.

3. Update Your Procfile #

Make sure your Procfile points to the script, like:

web: ./start.sh

Just remember: Heroku's filesystem is ephemeral, so config.txt will be recreated every time your dyno restarts—but this script will handle that automatically each time.

内容的提问来源于stack exchange，提问作者lambad