rfuehrer / syno_autoshutdown Goto Github PK

Add additional placeholder which replace #RUNLOOP_TIME# in human readable format. The placeholder #RUNLOOP_TIME# is in minutes.

have a look at the time values (6:50->7:43->40mins?). config is set to NOTIFY_ON_LONGRUN_EVERY=120, so the warning should be written after 2h and not after 40mins...

• wrong point of time
• wrong time calculation

12.11.2019 07:43:43 [18739] [I] - Sending notification (MESSAGE_LONGRUN)
12.11.2019 07:43:43 [18739] [D] - DS2413
12.11.2019 07:43:43 [18739] [D] - Das System ist bereits seit 12.11.2019 06:50:24 für 0d:0h:40m:0s online. Weiterhin erforderlich?

Although it is perfectly fine to use parameters as $1and $2 in the code like here...

Example

 notification() 
 { 
 	local system=$1    # myname
        local message=$2
 
 	if [ "x$IFTTT_KEY" != "x" ]; then 
 		if [ "x$IFTTT_EVENT" != "x" ]; then 
 			if [ "x$system" != "x" ]; then 
 				MY_NAME=$system 
 				MY_STATUS=$message
 				writelog "D" "$MY_NAME" 
 				writelog "D" "$MY_STATUS" 

If no system is names (or variable CHECKHOSTS is commented), the check found "0 systems" and the loop is resetted. The loop should say "No valid systems fund"

The brackets around the shell commands like in the example below serve no purpose and should be removed.

Every necessary variable should have a default value which is set if no config definition is set.

The code must be documented as comprehensively and meaningful as possible. Important lines/blocks in the code should be sufficiently commented and all functions should be provided with an appropriate header which describes the function, the input variables as well as the output variables.

Several instances of the script must be checked and handled correctly depending on the future scope of functions.

Starting the script from within itself requires at least several instances for a short time (2). Also, later parameterized calls to terminate all instances or other functions implemented in a single script could lead to several instances being (temporarily) executed at the same time.

Nevertheless, it should be considered whether the permanent execution of multiple instances makes sense. In particular, logging in different logs under different PIDs is not useful.

Wrong image embedded showing wrong command to start autoshutdown script.

The code is not consistent in how it invokes commands. Usage of backticks like in this example...

Notification messages should be enhanced with placeholders for dynamic content (list of found systems, elapsed time with no valid system etc.)

User data must be secured in such a way that the variables do not represent an entry gate for attackers but are also converted in such a way that the service can still be used.

Systems in standby or sleep state (e.g. Macbook) are detected as valid marker system and the system won't shutdown. Implement a fix to detect systems in standby and shutdown the NAS.

The shebang

This does not produce an error since on most Linux systems, sh is usually a symlink to bash or dash or some other "way more advanced then pure POSIX sh"-shell, but it's still wrong.

Fixing is easy: use #!/usr/bin/env bash as shebang.
(This also uses environment lookup).