[Bruno]

Hello,

Here is a (quickly made) PHP script that does a hard reboot on your rps if this one is down (ping OK but SSH down --> "read-only"). You have to run this script on a different server !

NB : it only concerns the RPS servers

The file hardRebootDoIfReadOnly.php :

/* This PHP script tests your RPS's status and do a reboot hard if it's down
Reboot hard only if your rps is in read-only */

############# begin config #############

    // OBLIGATOIRE
        $host = "rXXXXX.ovh.net" ;  // the rps name
        $login_manager = "XXXXXX-ovh" ;  //  your manager login...
        $passwd_manager = "*******" ;  // ...and its password
        $mail = "" ; //  the mail to send the hard reboot alert

    // OPTIONAL
        $reason = "OCO 398 Error read-only" ; // the reason of the hard reboot

############# end config ###############


// test the rps status with OCO (OvhChckOut) via telnet
    $test = shell_exec("telnet $host 79") ;
    $result = explode("\n", $test) ;
    $result = $result[3] ;

// get  the error code
    $code = intval($result[0].$result[1].$result[2]) ;

if ($code == 398)
{  // the RPS is KO, reboot hard !

    $soap = new SoapClient('http://www.ovh.com/soapi/ovh.wsdl');

    //login
    try {
    $login = $login_manager;
    $password = $passwd_manager ;

    $session = $soap->login($login,$password,null);

    echo "login successfull";

    } catch(SoapFault $fault) {
    echo "Error : ".$fault;
    }

    //dedicatedHardRebootDo
    try {

    $result = $soap->dedicatedHardRebootDo($session, $host, $reason, $mail, '');

    echo "dedicatedHardRebootDo successfull";

    } catch(SoapFault $fault) {
    echo "Error : ".$fault;
    }

    //logout
    try {

    $result = $soap->logout($session);

    echo "logout successfull";

    } catch(SoapFault $fault) {
    echo "Error : ".$fault;
    }

}

?>

The crontab directive (crontab -e) to test the rps every 10 minutes :

*/10 * * * * php /path/of/the/file/hardRebootDoIfReadOnly.php

You can modify it, here are the error codes returned by OCO :

#-- ERROR
300 - ERROR
301 - Connection time out
302 - Connection refused
310 - Login error
311 - Logout error
312 - Transfer error
350 - Time out of executing check script
398 - Result to old
399 - No result files