Question

The Code:

#!/usr/bin/php -q
<?php
while(true){
    $rate_limit = json_decode(file_get_contents('http://api.twitter.com/1/account/rate_limit_status.json'),true); # Check how many API calls remain

    $wait = round(3600/150);
    # *** IMPORTANT ***
    # Twitter limits API calls to 150/hr so regardless of the number of handles, we can only make one request every 24 seconds (which is 3600/150).
    # As a result, it takes one hour to run through 150 handles one time each.  Since we'll probably never monitor that many, each handle may get updated multiple times an hour.

    # Establish the database connection
    if(!$mysqli = mysqli_connect('localhost','twitterd','password','twitterd')){
        file_put_contents('tweet_gremlin.log', date('r') . ' Cannot connect to the database');
        break;
    }

    $get_handles = mysqli_query($mysqli,"SELECT `handle` FROM twitter_handles"); # Grab the Twitter handles from the database

    #  Append them to the $handles array
    while($row = mysqli_fetch_array($get_handles,MYSQLI_NUM)){
        $handles[] = $row[0];
    }

    #  Check the remaining API calls
    if($rate_limit['remaining_hits']<count($handles)){
        file_put_contents('tweet_gremlin.log', date('r') . ' Rate limit reached', FILE_APPEND);
        sleep(600);
        continue;
    }else{
        # Loop through the $handles values, make an API call, and insert the tweets.
        foreach($handles as $value){
            file_put_contents('tweet_gremlin.log', date('r') . ' Processing ' . $value . '\'s data...' . PHP_EOL, FILE_APPEND); # Tell the log what we're doing

            $user_data = json_decode(file_get_contents('http://search.twitter.com/search.json?q=from:' . $value . '&rpp=100&include_entities=1'),true); # Get the handle's timeline and put it into $user_data
            $user_data = $user_data['results']; # Put only the results index (tweets) into into $user_data

            if(count($user_data)<1){
                file_put_contents('tweet_gremlin.log', date('r') . ' No data for ' . $value . PHP_EOL, FILE_APPEND);
            }

            for($i=0;$i<count($user_data);$i++){
                # Lazy method for sanitizing variables
                $id = mysqli_real_escape_string($mysqli,$user_data[$i]['id']);
                $created_at = mysqli_real_escape_string($mysqli,$user_data[$i]['created_at']);
                $from_user_id = mysqli_real_escape_string($mysqli,$user_data[$i]['from_user_id']);
                $profile_image_url = mysqli_real_escape_string($mysqli,$user_data[$i]['profile_image_url']);
                $from_user = mysqli_real_escape_string($mysqli,$user_data[$i]['from_user']);
                $from_user_name = mysqli_real_escape_string($mysqli,$user_data[$i]['from_user_name']);
                $text = mysqli_real_escape_string($mysqli,$user_data[$i]['text']);

                $needles = array('hack','tango down','dump','breach','data');

                # Check the tweet relevance
                foreach($needles as $needle){
                    $needle = '/' . $needle . '/i'; # So we don't have to manually type out the regex

                    # Make a call based on the tweet contents
                    if(preg_match($needle,$text) == 0){
                        file_put_contents('tweet_gremlin.log', date('r') . ' No relevant tweet data in tweet #' . $id .PHP_EOL, FILE_APPEND);
                    }else{
                        file_put_contents('tweet_gremlin.log', date('r') . ' Processing tweet #' . $id .PHP_EOL, FILE_APPEND);
                        $insert_tweets = "INSERT INTO tweets (`id`,`created_at`,`from_user_id`,`profile_image`,`from_user`,`from_user_name`,`text`) VALUES ('{$id}','{$created_at}','{$from_user_id}','{$profile_image_url}','{$from_user}','{$from_user_name}','{$text}');";
                        mysqli_query($mysqli,$insert_tweets);
                    }
                }
            }
            file_put_contents('tweet_gremlin.log', date('r') . ' Sleeping for ' . $wait . ' seconds before processing the next handle' .PHP_EOL, FILE_APPEND);
            sleep($wait);
        }
    }
    mysqli_close($mysqli);
}
?>

If I run this on the command line like this:

./tweet_gremlin.php

It runs fine. But if I run it like this:

./tweet_gremlin.php &

It does nothing. I cannot figure out why.

EDIT: I had to take out pretty much all of the code because it wouldn’t let me submit it all.

Answer 1

You create a background job with the ampersand.

Depending on the underlying OS the background job changes to status “Stopped” because it wants to read from STDIN and/or wants to write to STDOUT.

Enter jobs into the same terminal after starting ./tweet_gremlin.php &
This tells your job’s status.

If the job’s status is “Stopped” discard STDOUT+STDERR and STDIN: ./tweet_gremlin.php </dev/null &>/dev/null &

If your script runs now you need to handle STDIN and/or STDOUT+STDERR differently in ./tweet_gremlin.php &
E.g. write to a logfile instead of STDOUT+STDERR.

You can also test the behavior of your terminal.
The following creates a PHP error on STDERR due to a missing semicolon:

php -r 'echo "backgrounding sucks\n"' &
jobs
fg1

The follwing creates normal output on STDOUT:

php -r 'echo "backgrounding sucks\n";' &
jobs
fg1

If you cannot fix the issue in your code, look into your terminal’s settings. Watch out for stty tostop (the SIGTTOU signal)