NotVeryLocalAdmin is a free to use server manager for SCP Secret Laboratory dedicated servers. It's built to replace LocalAdmin which ships with the dedicated server as a launcher and TCP listening server. The purpose of this project is to do almost everything LocalAdmin does but with extra features for babysitting your server(s).
NotVeryLocalAdmin supports daily restarts of which you can specify what time of day (locally) to restart all servers. It also features a memory check function which can check your systems free memory and restart your servers gracefully (either using restartnextround or using silentrestart) in the event there is a memory leak. It will also warn you if your memory usage is too high.
The primary feature of this program is a constant server check that will check to make sure the server console is alive and restart the process should it notice the server is not responding. We never thought we'd need a feature like this but Northwood x Exiled brought this up.
The best part, this program supports multiple servers instead of running multiple iterations of LocalAdmin. Neat.
If you want to just directly launch the program, your going to have to have Node.js installed. I built this project in version 15, but if you use older versions try to stay above version 12. If your lazy use nvm to switch node versions simply.
Using the node version is simple, you just need to run npm i first to install dependencies, then node index.js in the project folder and it will search the current working directory for the config file. Alternatively you can start the program with a config file argument to override this part of the program. Usage: node index.js -config <config.json path here>
For gods sake, please make sure you've CD'ed to the correct folder, otherwise you'll have errors and log files showing up places you don't want them..
Current versions of the project are packed using pkg and released on the github. There's a version for linux, windows, and macOS (if for some reason you manage to get a Dedi on a f***ing mac don't ask).
Using all these is simple, you just need to run the executable and it will search the current working directory for the config file. Alternatively you can start the program with a config file argument to override this part of the program. Usage: ./NotVeryLocalAdmin -config <config.json path here>
For gods sake, once again, please make sure you've CD'ed to the correct folder, otherwise you'll have errors and log files showing up places you don't want them..
The most important part is the config file. Your config file should be in your current working directory when you run the application, it should be called config.json and contain the following properties:
SCPExecutable: Should be a string that points directly to the SCP SL executable for the server. If windows SCPSL.exe, if linux SCPSL.x86_64. This could change if Northwood decides to change the executable name so check file paths.
restartTime: Should be an object containing two properties, hours, and minutes. Both are just numbers corresponding to the time to restart the servers automatically in local time. The time is in 24 hours, starting from 0-23. If you need to disable restarts, chose any negative number for hours.
checkinTime: Should be a number in seconds for how often to check the status of the server. This check is intended to watch for when the console freezes, and restart the server if the console fails to respond 3 times. The timeout however is 10 seconds. A config option for this may be added in the future pending requests.
serverStartTimeout: Should be a number in seconds, when the server is first started the program starts a timer which will wait this amount of time for the server to reach the ready state from start up. The ready state is typically when the server prints "waiting for players" in the console. Only when this event is fired does the timer stop. If the server takes too long, the program will restart the server assuming something went wrong.
serverRestartReqTimeout: Should be a number in seconds, this is the amount of time the server will wait for the server to respond to a restart request. Whether its restartnextround or silentrestart, the client will wait this time 3 times. If the server does not respond, it will forcibly restart the server by killing the process, assuming the process froze.
logFolder: Should be a string file path pointing to a folder to store log files in. This can be from current working directory, EX. ./Logs, or be somewhere completely different. In the future, the option to disable logging may be added.
minimizeLog: Should be a Boolean value (0 or 1 can also be used I suppose), this will toggle a log file cleaning feature I made. Its designed to clean up a few of the common errors or messages that SCP SL's STDOUT prints to help reduce the size of your 15GB log files. PLEASE, if your trying to debug something in this game, doesn't matter what it is, disable this feature, its possible it will clear debug lines or exceptions that your looking for in your logs.
logStdio: Should be a Boolean value, this controls whether or not to log the STDOUT and STDERR pipes of the SCP SL executable to the server logs. Please note STDERR is ALWAYS logged because you'll regret not capturing that one exception.
loggingMaxDays: Should be likely a string such as "30d", which will tell the log rotation to delete logs older than 30 days.
loggingMaxSize: Should be a string that tells the log rotation to rotate log files when they reach a certain size, example: "20m" is 20MB. Use 'k', 'm', or 'g' for this value.
memoryChecker: Should be a Boolean value, which controls whether or not to preform memory usage checking
respawnOnFullRestart: Should be a Boolean value, can be null, controls whether or not to respawn the NVLA process when fully restarting the program (this is used when you use the fullr or ffullr commands, before the program quits it will spawn a process of itself then exit). It is recommended to only use this if you don't have any kind of watchdog watching this process, note that you likely will lose control of the console entirely after a restart, so its best to not use this.
restartRate: Should be a string with a number and time suffix of either m (minutes), h (hours), d (days), mon (mons), or for some reason y (years). Should look something like "5h" for 5 hours. It DOES NOT support combining stuff like "5h2m". This will restart your servers every x amount of time gracefully, it will override the daily scheduled restart. This timer is based on the moment NVLA started, NOT the moment individual servers started. This can be set to Boolean false to disable it.
servers: Should be an array [], of objects. Each object should be a server with the following properties:
"uid": "1234abcd", //This should be a unique Identifier that the program can use to differentiate this server if theres servers with the same labels. No two servers should have the same uid
"p": 7780, //The port number of this server
"l": "Server-4", //This is a server label, its cleaner than a UID, gives it a nice title which is used in the console and logging.
"disabled": true //Whether or not this server is disabled, if disabled, it will not auto start, and cannot be started until enabled
In case you need it, here's an example config for 5 servers on different ports:
{
"SCPExecutable": "A:/SteamLibrary/SteamApps/common/SCP Secret Laboratory Dedicated Server/SCPSL.exe",
"restartTime": {
"hours": 4,
"minutes": 0
},
"checkinTime": 5,
"serverStartTimeout": 60,
"serverRestartReqTimeout": 3,
"logFolder": "./Logs",
"minimizeLog": true,
"logStdio": true,
"loggingMaxDays": "30d",
"loggingMaxSize": "20m",
"memoryChecker": true,
"restartRate": false,
"servers": [
{
"uid": "12345a",
"p": 7777,
"l": "Server-1",
"disabled": false
},
{
"uid": "12345b",
"p": 7778,
"l": "Server-2",
"disabled": false
},
{
"uid": "12345c",
"p": 7779,
"l": "Server-3",
"disabled": false
},
{
"uid": "12345d",
"p": 7780,
"l": "Server-4",
"disabled": false
},
{
"uid": "12345e",
"p": 7781,
"l": "Server-5",
"disabled": false
}
]
}
Note: You can also opt to specify your config file somewhere else instead if you wish. Use -config <path> for arguments when starting the program
The console in this program also has some commands, note that the console can be a little janky in SSH sessions:
stop - (Usage: stop <server label|UID|port>) Stop a specific server
start - (Usage: start <server label|UID|port>) Start a specific server
restart - (Usage: restart <server label|UID|port>) Restart a specific server restartforce | rf | fr - (Usage: fr <server label|UID|port>) Forcibly restarts a specific server, first with sr, then with killing the process if that doesn't work. Future commands might be added for literally deleting the server process.
enable - (Usage: enable <server label|UID|port>) Enable a specific server, saves it to the config too
disable - (Usage: disable <server label|UID|port>) Disable a specific server, saves it to the config too
exec | run - (Usage: run <server label|UID|port> ) Sends a console command to the specific server. Everything after the server Identifier is sent to the console as a string so it can be any length
quit | exit - (Usage: quit) Stops all servers and exits.
startAll | sa - (Usage: sa) Starts all the servers that are enabled
stopAll | sta - (Usage: sta) Stops all the servers that are running
restartAll | ra - (Usage: ra) Restarts all the servers that are running or not running (if they are enabled)
enableAll | ea - (Usage: ea) Enables all configured servers, saves to config
disableAll | da - (Usage: da) Disables all configured servers, saves to config
version | v - (Usage: v) prints the current NVLA version
fullRestart | fullr - (Usage: fullr) Will close the program when all servers are considered "inactive" (all offline or all empty). Optional: you can make it respawn the process on exit with respawnOnFullRestart in your config file set to true.
forceFullRestart | ffullr - (Usage: ffullr) Will close the program immediately but it will use silent restart, which means players will get a round restart event instead of server closed. This means you can fully close everything and then restart this program immediately while still catching all the players you disconnected. 'fullRestart' does the same but it will just wait until its safe to do so. Optional: you can make it respawn the process on exit with respawnOnFullRestart in your config file set to true.
reload - (Usage: reload) Reloads the whole config file and reconfigures the program live. Servers will not be affected by this function and the function will update log folder paths and such. Servers that are added will be created, servers that are removed will not be deleted until the program is restarted as a safety measure.
In the future I might make the program a bit more optimized in terms of handling current working directory and the actual current directory the program is running in. Some other config options may also need to be added to make things a little bit more configurable for hosts but I'll add what people request.
If you have any issues feel free to post them. If you run into errors or crashes please provide me with screenshots or console output as well as what you did to reach that state. Please note this repo will only be maintained as often as I can be available. I have a life too, and I can't spend it babysitting servers, let alone repos, why do you think I made this? :kek:
Its generally a really good idea to use a good watchdog process that can watch this program and restart it if anything happens to it, say a memory leak or a crash. They happen and sometimes you need to deal with it. I recommend using something like systemctl or making this program a service in your operating system so your OS can take care of it. You have a lot of options and personally if your on Linux I recommend using a combination of screen + systemctl, where you make systemctl launch a screen of the program. Theres a special unit file config you can use for launching the screen and capturing the forked process PID so systemctl knows which pid to watch.
If at some point your program crashes, I setup the program to catch any major exceptions and dump them into a crashLog.txt file in the current working directory. If you pass this to me it would help a lot in catching some of the bugs that I may have missed in putting this together.
If you want to help with this project you may contact me at alex1001(at)live(dot)ca Or via discord if you know where to find me. Preferably discord cause my email is a trash can.
If your on windows, fair warning about a strange bug with the packed version of this program, if you kill the program using the X button on the window, it tells all the dedicated servers to exit which causes them to quit Unexpectedly which causes the program to restart them while windows is trying to kill the program. It behaves very strangely and I can't figure out how to have node capture when this happens. Please use the quit or exit commands instead when exiting.. If you DON'T want to do that and continue using the X button, don't complain to me why this happens cause I have no clue and I did my best to fix it ¯_(ツ)_/¯