archive-sharding

[fiery hornet]

[SHARDING_READY_TIMEOUT]: Shard 0's Client took too long to load

I got this error a long while back and I don't remember how I fixed it. Now I got it again and I'm not sure what to do

Here's my sharding file

const config = require('./config');
const { ShardingManager } = require('discord.js');

const manager = new ShardingManager('./shard.js', {
    totalShards: 'auto',
    token: config.token
});

// Trying to spawn the required shards.
manager.on("shardCreate", shard => {
    shard.on("ready", () => {
        console.log(`[DEBUG/SHARD] Shard ${shard.id} connected to Discord's Gateway.`)

        shard.send({type: "shardId", data: {shardId: shard.id}});
    });
});

manager.spawn().catch(error => console.error(`[ERROR/SHARD] Shard failed to spawn ${error}`));

[little bluff]

So I'm in 40K servers and have been using external sharding - or sharding manager.

From what I've heard, the primary benefit of using external sharding is memory usage. What are the other benefits, if any?

The reason I ask is, 40K servers means 40 shards, which means:
• 40 MySQL connection pools — each of which reserves a huge chunk of memory.
• 40× cached resources — such as images used for rendering.
• Etc.

My VDS sits at 1-2% CPU. However, I've consistently had to upgrade because of memory, mostly thanks to MySQL. I've recently integrated slash commands, which has completely eliminated my need for GUILD_MESSAGES, so I no longer have to test each message against ~25 commands. This will likely reduce my usage even further. I don't cache users, messages, anything, and fetch it all when needed.

Can I get away with internal sharding?

[lusty smelt]

How much is the maximum amount of total guild count internal sharding can handle?

Its likely ideal upto 32K i think

[little bluff]

What would limit its ability to go higher?

[lusty smelt]

Happened to see this https://www.npmjs.com/discord-hybrid-sharding
Utilising internal sharding and external together, it claims to reduce memory usage

[lusty smelt]

I think you cant move around servers in case one process goes down. Or is that for clusters? Not sure

[lyric lantern]

I want to send an embed between shards and I made it like this:
channel.send({ embed: JSON.parse('${JSON.stringify(messageToSend)}') });

But I got this error:

UnhandledPromiseRejectionWarning: SyntaxError: missing ) after argument list
and points to discord.js/src/client/Client.js:425:17

What did I do wrong?

[crimson pond]

You are missing ) somewhere in that line

[lyric lantern]

I know that it derivated it means that, but I double checked it and there's no missing )s in this part:

channel.send({ embed: JSON.parse('${JSON.stringify(messageToSend)}') });
and If I try simply:
channel.send(messageToSend);
It's output is [Object object]

[lusty smelt]

[14/05 07:24:56][LOG] [WS => Shard 2] [HeartbeatTimer] Didn't receive a heartbeat ack last time, assuming zombie connection. Destroying and reconnecting.
Status : READY
Sequence : 396778
Connection State: OPEN
[14/05 07:24:56][LOG] [WS => Shard 2] [DESTROY]
Close Code : 4009
Reset : true
Emit DESTROYED: true
[14/05 07:24:56][LOG] [WS => Shard 2] Clearing the heartbeat interval.

Why doesnt the shard reconnect after destroying

[crimson pond]

to use template literals you will need to use ` not '
you probably made a typo on that

[next plaza]

since your cpu usage is very less, you can combine internal sharding with sharding manager, which will save a lot of ram. With the combination you would need 8 shards and 5 internal shards on each shard (would be the 1/5 ram usage of your current one), as Joshua suggested, you can do it with the package. It combines both or you have to make your own solution

[wet topaz]

1/5? that doesn't sound right

[lyric lantern]

I use ` on the start and end of the template

[crimson pond]

But in the code you sent you used '

[fervent zealot]

.

[pallid spindle]

Hi, how does sharding work?

My bot uses another class that extends the discord.js client

That class has a function that loads everything and logs in the bot

Will sharding work when I call this function in the main file? (slowmode is a bit annoying)

[next plaza]

Sharding spawns every time new processes, 50mb is node usage for one process. He spawns 40 shards means 2gb, on 8 shards it would just be 400mb

[young totem]

How can I make a rolling restart on my shards?
To clarify, restart one shard at a time to minimize downtime when making changes

[wet topaz]

well, if you specifically mean overhead of sharding alone, sure

but if you're talking about overall size, caches won't play as nicely

---

you could simply broadcastEval a process.exit

using shard id

[young totem]

so that would make that sharding client restart that shard? when it detects it went offline?

[wet topaz]

sharding manager restarts processes when they exit, yes

[young totem]

sounds good, ty!

[terse cairn]

const data = require('./data.json')
const { ShardingManager } = require('discord.js');
const manager = new ShardingManager('./index.js', { token: data.token, respawn: true, totalShards: 2 });
manager.on('shardCreate', shard => {
  console.log(`Launched shard ${shard.id}`);
  shard.on("ready", () => {
    try {
      shard.channels.cache.get("789294581333884948").send(`Shard ${shard.id} preparada.`)
      } catch {
        console.log("ERROR");
      }
  })
  });
manager.spawn();

it never spawn the shards

[terse cairn]

?sharding

[icy crow]

const { ShardingManager } = require('discord.js');
const Manager = new ShardingManager('./bot.js', { token: 'your-token' });
Manager.spawn();

https://discord.js.org/#/docs/main/stable/class/ShardingManager
For an in-depth guide: https://discordjs.guide/sharding/

[wet topaz]

that... is not what shard is

also, are you getting any errors? anything?

can't pull out answer out of thin air

[terse cairn]

nope, it doesn't throw any error
just doesn't work

[wet topaz]

what are you running

[terse cairn]

rn, this:

const data = require('./data.json');
const { ShardingManager } = require('discord.js');
const manager = new ShardingManager('./index.js', { token: data.token, respawn: true, totalShards: 3 });
manager.on('shardCreate', shard => {
  console.log(`Launched shard ${shard.id}`);
  shard.on("ready", () => {
    try {
      shard.channels.cache.get("789294581333884948").send(`Shard ${shard.id} preparada.`)
      } catch {
        console.log("ERROR");
      }
  })
  });
  manager.createShard(0);
  manager.createShard(1);
  manager.createShard(2)

i've tried with spawn and didn't work so now i'm trying with createShard

[wet topaz]

1. this is not how that works. shard isn't client
2. use spawn() instead of calling createShard n times

what command are you using to start

[terse cairn]

a command to start..?

[wet topaz]

yes

that code doesn't run on its own

you have to start it

[terse cairn]

OHHH so that's the problem, i don't have any command to run it

[wet topaz]

wat

did you even attempt to start your bot

[terse cairn]

yes i did.

[wet topaz]

how

[terse cairn]

the bot runs perfectly, the problem is that the shards doesn't get created

[wet topaz]

how do you start it

[terse cairn]

with the button "start", i use replit

[wet topaz]

fairly sure that that button starts main file

which in 99.9% of cases is index.js

[terse cairn]

mmm

[next plaza]

you have to go to the .replit file and change 'run=node shard.js'

[terse cairn]

ok now, it works, but, now express throws error, 'Port 3000 is already on use'

[broken sequoia]

not djs

[terse cairn]

it's djs, i'm trying to init sharding and it doesn't create the shards correctly because the express throws error

Error: listen EADDRINUSE: address already in use :::3000

(node:1005) UnhandledPromiseRejectionWarning: Error [SHARDING_READY_DIED]: Shard 0's process exited before its Client became ready.

[terse cairn]

someone?
:c

[jade venture]

Is there way to check if all shards has spawned?

[wet topaz]

@terse cairndon't run the server in index.js but in manager file.

[jade venture]

Is it possible to "build" Client from array? (Using broadcastEval)

[terse cairn]

Ok

[terse cairn]

just after shardCreate, i try to use .eval but it says Shard 0 has no active process or worker

is there event for when worker is launched

[wet topaz]

code?

[terse cairn]

manager.on('shardCreate', async shard => {
    console.log(`Launched shard #${shard.id}`);

    if(commandList.length === 0) {
        commandList = await shard.eval('Object.keys(Commands)');
        console.log(commandList);   
    }
});

https://lune.dimden.dev/6328cc045a.png

[wet topaz]

fairly sure that that event is emitted when the shard is created, not when it became ready. for that there is Client's ready event

[terse cairn]

yeah but i need to eval things

i don't care what's going on inside shard

[wet topaz]

then eval them in ready

[terse cairn]

uhmm how

[wet topaz]

?

[terse cairn]

how do i get client ready event if it's "inside" shard

[wet topaz]

the regular ready event on the client

in the file your manager spawns

[terse cairn]

i have to do it outside, not inside bot

[wet topaz]

why exactly?

[terse cairn]

because of how bot built

[wet topaz]

then make it not be built like that and use ready event on the client, which is de facto the thing that uses commands

manager has absolutely no correlation to messages whatsoever

[terse cairn]

i don't need messages or anything inside, just some thing

i'm hosting web api for it outside

[wet topaz]

from what i can see above, all you do is log commands

[terse cairn]

i can't make web server run 50 times inside every shard

[wet topaz]

not sure what is the difference between getting them form manager or from a shard

then don't

run server along the manager

and connect to the server from shard

shard -> manager's server

not manager's server -> shard

[terse cairn]

i don't quite get it

[wet topaz]

you know how websites work?

there is one server

and every browser connects to that server

as a client

it doesn't work in a way that your browser runs a local server and waits for the website to connect to your browser

same logic here. Run one server, in the sharding manager file

and then in ready event in shards, connect to that server and send whatever you need

[terse cairn]

i have to for example change some settings of bot in guild via api (oauth2)
and i'll have to do server -> shard not shard -> server

client makes request to server, server does things to shard where server is

and the question was just to get some variable from shard
immediately after it runs

[wet topaz]

again, get that variable in ready

there are tons of ways you can maintain a two way connection between manager and shards

websockets, pub-sub or whatever you want

but the general idea is to not run a server everywhere

[terse cairn]

so you tell me to make something like webhook or websocket server and then the shard will connect to it and send info via that
i think that's 50 times harder than just evalling from sharding file

[wet topaz]

no, not really

[terse cairn]

i just need to get 1 variable from the first shard that will connect

[wet topaz]

#archive-sharding message

[terse cairn]

#archive-sharding message

[wet topaz]

#archive-sharding message

are we really going to do that

[terse cairn]

lol, i just don't think it's necessary

a harder way of doing things

[wet topaz]

not harder, but proper

it's really not hard to run a single websocket server

[terse cairn]

what's wrong with evalling things with sharding manager

[wet topaz]

why are you even evaling command list

[terse cairn]

because i need it for stuf

[wet topaz]

then read the directory

[terse cairn]

commands are inside the module files

[wet topaz]

then read the files

[terse cairn]

mfw running huge ass slow files that take 10 seconds to run just to get command list

[wet topaz]

you do it 50 times anyway

[terse cairn]

it was an example
and still, bad thing

why running for no reason

[wet topaz]

ok then

go ahead and use your eval

[terse cairn]

(it doesn't work which was the original issue)

[wet topaz]

and any other monkey patched solution

see

this is why i suggested multitude of things

[terse cairn]

but why it calls create event before it's created??

[wet topaz]

but you just said "what was wrong with my way that literally didn't work"

because that it's the event lol

it's shardCreate

not shardReady

for that tiere is client ready

literally

[terse cairn]

it would work perfectly if i got event smething like "shardLoaded"

[wet topaz]

it's called ready in client

[terse cairn]

gsjhgsdjkhd, yeah whatever

[wet topaz]

also, wrong

[terse cairn]

https://lune.dimden.dev/e42a11625e.png

[wet topaz]

yes

because

and only later it awaits all promises from spawning

[terse cairn]

everything would be 5 times easier if event like that existed

[wet topaz]

dude

why are you so opposed to using ready event

[terse cairn]

because have to make a server just for 1 (one) thing

[wet topaz]

no, you don't have to

[terse cairn]

wait is there way of communicating with sharding manager

[wet topaz]

you can even console log it

and read from stdout

and yes

[terse cairn]

gg

i have 549045903 of different stuff in console log

[wet topaz]

and === or startsWith() don't really care

[terse cairn]

i guess i'll prob have to remove almost all of console logs

i'm just about to hit 2500 servers and working on huge update with web dashboard and stuff

so how to send data to sharding manager inside shard?

k found

[winged bobcat]

You could have listened to the shard ready event

That gets emitted after the client

[mental tinsel]

    at Timeout.onTimeout (/home/runner/Amity-Bot/node_modules/discord.js/src/sharding/Shard.js:163:16)
    at listOnTimeout (internal/timers.js:554:17)
    at processTimers (internal/timers.js:497:7)
(node:229) UnhandledPromiseRejectionWarning: Unhandled promise rejection. This error originated either by throwing inside of an async function without a catch block, or by rejecting a promise which was not handled with .catch(). To terminate the node process on unhandled promise rejection, use the CLI flag `--unhandled-rejections=strict` (see https://nodejs.org/api/cli.html#cli_unhandled_rejections_mode). (rejection id: 1)
(node:229) [DEP0018] DeprecationWarning: Unhandled promise rejections are deprecated. In the future, promise rejections that are not handled will terminate the Node.js process with a non-zero exit code.```

How would I handle this error?

[winged bobcat]

?shardreadytimeout

[icy crow]

Got this error: Error [SHARDING_READY_TIMEOUT]: Shard 0's Client took too long to become ready.?
The spawnTimeout can be set to -1 or Infinity to prevent future errors:
https://discord.js.org/#/docs/main/stable/class/ShardingManager?scrollTo=spawn. This can also happen if you call ShardingManager#spawn before attaching a listener to the shardCreate event which could create a race condition possibly preventing the shard 0 to perform a successful launch.

[teal plume]

Could someone send me an initial bot for me? to train some things without a bugs

[wet topaz]

?guide

[icy crow]

The official discord.js guide, maintained by core members of the community https://discordjs.guide/

[terse cairn]

Quick question: if my bot sends a message with a custom emoji, there's no need for me to use use broadcastEval to pull the id and name from cache if I alraedy know the id and name. I could just message.channel.send() with the <:name:id> as is. Am I right? I did this and it appears to work, but I just want to be sure. To clarify, I'm simply hard-coding the emoji name and id.

No you don't need broadcast eval i think

[sturdy ember]

how do I get all the users in all the shards?
it should be like this right?

${results.reduce((acc, userCount) => acc + userCount, 0)}

because to show all the servers it is like that

ping me with the anwser pls

[wet topaz]

You could always try

[terse cairn]

how to get the shard ping?

[next plaza]

users are just cached, when you fetch or process them, so you would not get all users, you have to rathe reduce the membercount of the guild

[next plaza]

#client.ws.ping

[terse cairn]

this is the shard ping?

[next plaza]

the shard ping, is the ping of the client, so you have to broadcastEval this.ws.ping

[terse cairn]

ok

[stiff mason]

what would cause this

[marsh finch]

As the string mentions, Discord wants you to reconnect every (15 min - 4h) to balance the load on their end.

[stiff mason]

oh

[minor sapphire]

I was wondering about shards. My bot isn't completely huge, but what is the benefit of shards, and what are they used for?

[dense cape]

Sharding splits ur bot into separate processes

[golden moss]

You wanna use it before you hit 2500 because its required by discord at that point.

[dense cape]

Discord requires you to do it at 2500 guilds

It distributes the load of events across how ever many shards you have

[golden moss]

Is there a way to get the whole GuildMemberRoleManager#cache from broadcastEval? @dense cape

[dense cape]

You can, but it would be a plain object

[golden moss]

Would I not be able to call .get()

[dense cape]

No

Collection would be converted to an array of arrays

[golden moss]

hmmm

[dense cape]

What are you trying to achieve?

[golden moss]

Getting a set of roles from a guild via broadcastEval then I run a for loop through my set of roles in my DB and attempt to call .get() and then if the role is found display it on my dashboard for a role select dropdown menu

[dense cape]

So you need to list the roles, then be able to get a role on command?

[golden moss]

Yes

[dense cape]

Are these actions separate endpoints?

[golden moss]

No

[dense cape]

You could just throw the array of arrays back into new Collection/Map, then get the role data

[golden moss]

ohhhh

Hmmm I am wondering how the best way to do that is

Wouldn't it just be new Collection(data)? or should I run a loop to put it all in?

[dense cape]

The former

[golden moss]

Why do I not know what former means all of a sudden in this context

[broken sequoia]

First

[golden moss]

Ah

Finally, I have been working on a fix for forever and you helped so much @dense cape

Well, I am now having a issue where the collectionn isn't actually inputting any data values

[dense cape]

Show code

[golden moss]

let roles = await client.shard.broadcastEval(`this.guilds.cache.find(g => g.id === '${req.params.serverID}').roles.cache.toJSON()`)
            console.log(roles)
        roles = new Collection(roles)

http://img.extreme-is.me/nA3onRrazclmDDuu (I logged both of my 2 collections at bottom)

[dense cape]

broadcastEval still returns an array of data

[golden moss]

Oh wait

[dense cape]

One item for each shard return data

[golden moss]

But it would only return one set so I need to do roles[0] when I pass it

[dense cape]

Yea

[golden moss]

Lets see

Both collections still empty?

http://img.extreme-is.me/vyG7uCvRY7EqlDW1

console.log(roles[0])
    roles = new Collection(roles[0])
console.log(roles)

@dense cape

[dense cape]

Huh, it returned an array

You could do [...<Map>] in the eval

[golden moss]

Where?

[dense cape]

In the eval

Instead of calling toJSON, wrap the Collection in an array and use the spread operator

[golden moss]

I don't follow

[golden moss]

How should I wrap it?

remove the .toJSON() and switch it with?

[dense cape]

Guild.roles.cache is a Collection (extends Map)

[golden moss]

Yes

[dense cape]

If you do [...<Map>], it’ll return an array of entries [key, value] that can be easily converted back to a Map/Collection

[golden moss]

Why is my brain breaking

[dense cape]

[...<Guild>.roles.cache]

[golden moss]

ahhhh

[dense cape]

You would have to add a check to make sure the guild is found tho

[golden moss]

Yeah I already do

I just gotta figure out how to do this because my brain is hurting for some reason, I am for some reason thinking I am supposed to do the [...roles.cache] inside the broadcastEval, I can't understand jack shit rn

[dense cape]

Yes

[golden moss]

Is that right?

[pallid depot]

Error [SHARDING_READY_DIED]: Shard 0's process exited before its Client became ready.

[tawny silo]

`same error ;/

[pallid depot]

I put manager.respawnAll() in a setInterval and that error is gone

[tawny silo]

oh thanks for the help

[sudden sequoia]

did the tags solution not fix it?

[tawny silo]

nope

[sudden sequoia]

hmm, try that i guess

[tawny silo]

there's no option coming like that

[sudden sequoia]

can you show your sharding file?

[tawny silo]

yes

const {ShardingManager} = require("discord.js");
require("dotenv").config()

const shards = new ShardingManager("./index.js", {
    token: process.env.TOKEN,
    totalShards: "auto"
});

shards.on("shardCreate", shard => {
    console.log(`[${new Date().toString().split(" ", 5).join(" ")}] Some shard is ready lol #${shard.id}`);
});

shards.spawn(shards.totalShards, 10000);```

[sudden sequoia]

wait a min

[pallid depot]

Seems clean for me

[tawny silo]

tf ,thank you i guess?

[tawny silo]

sure waiting

[sudden sequoia]

shards.spawn(1, 1000, -1);

can you try this?

[tawny silo]

sure

[sudden sequoia]

ping me with the response 👍

[tawny silo]

no response

its just stuck

[tawny silo]

nothing's happening its just stuck

[sudden sequoia]

hmmm

¯_(ツ)_/¯

[tawny silo]

...

[sudden sequoia]

now i am out of ideas why, you would have to wait until some one who knows about this comes

[tawny silo]

sure

[wet topaz]

does it spawn shards at all?
if so, did you try listening to debug event on the client?
eventually, do you do some tasks in client file that could probably disallow it from readying up?

[terse cairn]

His question has been answered in a different channel, he has to wait out a 15 hour too-many-logins cool down

[wet topaz]

i love channel hoppers.

[terse cairn]

why does my bot respond two times

nvm i fixed it

[thin trench]

Is there a way to have per shard status? So each shard has their shard number in their status

[dense cape]

?docs ClientUser.setActivity

[dense cape]

Use options.shardID

[pallid depot]

How to define shard in other files

[lean scarab]

?help

[lean scarab]

?docs Intents

[wet topaz]

#app-commands

[graceful juniper]

hello , who can help me with sharding? ty

[wet topaz]

guide can

[high sequoia]

whats the benefit of sharding?

[wet topaz]

splitting load that is eventually enforced on you

[high sequoia]

so bot goes faster?

[wet topaz]

no

not raelly

it more prevents it from slowing down, since one shard only handles some of the guilds

[graceful juniper]

who?

[wet topaz]

guide

https://discordjs.guide/sharding/

[graceful juniper]

I saw it, but still need help

[wet topaz]

well, if you actually ask, someone might help

note how you pretty much wasted 9 hours because all you asked was "who can help me", where in reality such question is pointless

[graceful juniper]

In those 9 hours I slept xD

[jagged fjord]

i was acting stoopid for like idk a long time

[wet topaz]

i know that you slept

but if you would ask actual question, you would have a chance to get up to answer

and not to need to ask, and then wait for any answer again

which you btw still not do

so you're stalling getting any answers

[graceful juniper]

Okay, need to do a shard for each command ? or just the main file of the bot

Do not really understand in shard

[wet topaz]

sharding spawns copies of your main file

your main file handles commands and events and all that

[graceful juniper]

so i need to this and that's it? ( i have commands handler)

const { ShardingManager } = require('discord.js');
const manager = new ShardingManager('./bot.js', { token: 'your-token-goes-here' });

manager.on('shardCreate', shard => console.log(`Launched shard ${shard.id}`));
manager.spawn();

[wet topaz]

yes

[graceful juniper]

ok , Thank you very much

[next plaza]

it is not recommanded to spawn manually new shards, use the method #manager.spawn(shard_count, spawn_delay, timeout)

yes, so whats your question?

[mortal matrix]

How does one get information on their bots shards? (Viewing how many guilds in said shard, viewing if shards are down etc etc)

[potent tendon]

See the pic, it's in bot.js and I have the following code in text in the server.js file, how can I make it work? Btw, the server.js is the file which starts the bot in package.json.

const manager = new ShardingManager("./bot.js", {
    token: "",
    totalShards: 2 
    });
manager.spawn();
const uptime = bot.channels.cache.find(channel => channel.id === "835884721463623721");
                                
manager.on("shardCreate", shard =>     console.log("Shard Number #${shard id} is now online!"));

manager.on("shardCreate", shard =>
    uptime.send("The Shard ${shard id} is **now** online and running the smoothly!"))```
Error: bot is not defined

[broken sequoia]

Yes, it isn't defined in the sharding file

[potent tendon]

So how can I make it work?

What should I add in the sharding file?

Or what can/should change?

[mortal matrix]

How does one get information on their bots shards? (Viewing how many guilds in said shard, viewing if shards are down etc etc)

[waxen hill]

Shard 0's process exited before it's Client became ready.

Why does this happen?
My bot has been fine since now.

Ah I found the issue.

Just a bit of code I added was stopping my client from logging in.

[potent tendon]

I know, my main file has code in aoi.js, see the msg above pic.

[pallid depot]

https://github.com/discordjs/guide/blob/master/code-samples/sharding/getting-started/12/bot.js

[graceful juniper]

I nees to start the shard file and not the main file ?

[pallid depot]

Yes

[graceful juniper]

Ok

All shards ? @terse cairn

[pallid depot]

In this link @terse cairn

[graceful juniper]

client.shard.count

Ah

[potent tendon]

If I keep multiple shards(0,1,2) while having 512mbram and 1GBssd, make the bot ping high?

For some reason, I just added the shards and the ping is 700+

What is it? I can try to.

[graceful juniper]

What i need to put here ?
In the "()"
manager.spawn();

[pallid depot]

https://discordjs.guide/sharding/#broadcasteval follow this

[graceful juniper]

For me to

[potent tendon]

Hmmmmmmmmmmm

Maybe the global database might be lagging? There might be a lag spike?

Or maybe did our host database had a lag spike? Not sure if our host is same.

[pallid depot]

[graceful juniper]

Anyone know why I get a high ping in shard?

[pallid depot]

On host or selfhost ?

[graceful juniper]

Host

But , i try 2 other hosting , and same ping ( high ping )

[pallid depot]

Maybe you didn't choose the right server !

[graceful juniper]

Believe me I chose the right server lol

[prisma jewel]

I believe its in shardingmanager's docs somewhere

@terse cairn Have you got any idea of how to get the bot to start sharding? Having trouble with it myself

when I try to start my script, looking very similar to this, it crashes my VPS and throws a lot of errors

client.shard.id

[graceful juniper]

What i need to put here ?
In the "()"
manager.spawn();

[prisma jewel]

manager.spawn([ShardCount], [timedelay], [timeout])

Yeah, those are all 3 params

?docs ShardingManager

[graceful juniper]

And anyone know why I get a high ping with shard?

[prisma jewel]

Sharding should not be used unless you either
A. Have a music bot
B. 1k+ guilds

it can actually hinder performance normally sometimes

[graceful juniper]

My bot is on 740+ servers, and there are bots on about 100 servers that do shard and nothing happens

[prisma jewel]

Are you using commands with sharding in mind?

sharding does use excessive amounts of memory

I'd recommend a VPS provider with an unmetered plan.

[iron solstice]

how many guilds is the recommended amount for music bots? I thought it was 1k guilds for everything?

[next plaza]

DIscord recommands 1k, but you can go higher, up to 2000 (2500 is the max)

[iron solstice]

i know that, but how is a music bot any different?

[strange echo]

It's up to how much memory each shard is gonna need based on what you're doing. Audio stuff has higher demands than text

[next plaza]

How much data do you process? could you check the ram usage of external buffers, log out process.memoryUsage()

[iron solstice]

oh ok, so is it recommended for music bots to shard earlier?

[potent tendon]

Is
manager.spawn([0,1,2], [25000], [300000]) the correct Usage? And it's supposed to start the shards 0,1 and 2 with a interval of 25s?

[wet topaz]

not sure why would you want to spawn them 25s apart, but still

if you want to spawn only those 3 shards, pass them as options to ShardingManager

spawn() doesn't take such thing at all

[potent tendon]

Okay.

And what's wrong here?

[graceful juniper]

anyone know why I get a high ping with shard?

[next plaza]

your client did not got ready, you can use a option in the manager.spawn, where you wait until the client gets readt

https://media.discordapp.net/attachments/824410765012238366/844219403574444092/unknown.png?width=617&height=333
Its the spawnTimeout

[white crane]

Hi guys, what would the limiting factors for the size of internal sharding be? Is there an established estimated limit to it?

[dense cape]

Ur cpu would be the main limiting factor

Bc internal sharding runs multiple ws on one cpu/process

[young totem]

Can I run code in ShardingManager simlar to with broadcastEval?

[young totem]

Sorry to bump but it's been a while. Anyone able to help?

[wet topaz]

what do you even mean

[young totem]

from a shard, run code with access to the shardingmanager and its functions/properties

in the parent process

[wet topaz]

what are you trying to do

[young totem]

run code to restart shards individually, including the one I initiated it from. I can’t run it on that shard since i’m stopping the process so I want to run it on the parent process (sharding manager)

[wet topaz]

there is https://discord.js.org/#/docs/main/stable/class/ShardClientUtil?scrollTo=respawnAll

[young totem]

yes but I want to do it one at a time so I can’t use restartall

[prisma jewel]

I shard my music bot at 72 guilds, that should give you a good idea of just how much memory is used

I have a question. So how would I pass data from my database into the separate shards

[dense cape]

Ur shards should be fetching from the db

@prisma jewel

[tropic rose]

the ready event is only fired when all shards are ready, right?

[marsh finch]

All shards handled by the emitting client, clients in other processes / workers do not count here.

[lyric lantern]

Do I see correctly that you are using a second bot account to check shard status?

good idea, thank you

back to your question: it may be a server provider issue

[next plaza]

thats normal that internal shards reconnect and resume, it can be bc of the internet connection or sometimes DIscord closes the connection.....

welcome

[prisma jewel]

How exactly would that be done

[potent tendon]

A cluster means a ShardingManager correct?

What responds with the guild count in all the shards?

And how can get all the guilds from all shards?

And what about all members from all shards?

[terse cairn]

How can i map all of the shards?

[prisma jewel]

ShardingManager.shards.forEach I think

[next plaza]

broadcast client.guilds.cache
same with members

[prisma jewel]

@next plaza Does .fetchClientValues() work with something like this

[next plaza]

when you load it on the client, yes

[prisma jewel]

thats in the client's module.exports, so I assume so

and it extends discord.client

So @next plaza this should work?

[prisma jewel]

I'd use client.shard.fetchClientValues('guilds.cache.get'), with the only issue being that it returns an array, and a promise, so you have to use await or .then()

[next plaza]

try and see xD

[next plaza]

and all properties will not be there

[prisma jewel]

Ah if you're trying to pull properties than your idea is better

[next plaza]

on broadcastEval it is the same xD

[prisma jewel]

damn discord.js is not making sharding easy lol

I feel like I may be about to hit a 429 because I forgot where my heartbeat setter is and I can't set it's heartbeat to 4 minutes instead of 20 seconds

[prisma jewel]

it definitely did not

[next plaza]

you can't change internal stuff

you can inject some code and overwrite properties, but this will break the libary

[prisma jewel]

so how can I pull data from the database across shards

[next plaza]

why dont you fetch it from the current shard?

[prisma jewel]

does it have database instances?

[next plaza]

you can create one???

[prisma jewel]

i have a database already

[next plaza]

yes, you can run the database an all shards

on the bot instance itself

[prisma jewel]

Alright, thanks

[young totem]

Never got a full answer on this, can someone help?

[prisma jewel]

shard.respawn(shardID)

?docs shard

[young totem]

I already know how to do that. As I said before, I need to access the parent process from the shard and run the code from there because the shard process that initiated it is stopping.

[prisma jewel]

I don't think that is possible. There is no real parent process, the shards are independent I thought

[young totem]

Isn't the sharding manager the parent process? If there isn't a parent process then how would the shards get respawned if the close?

[prisma jewel]

oh yes, the sharding manager is the parent. You'd have to call that in your shard.js file somewhere

I forgot that shardingmanager existed lol

[young totem]

Yes, I'm trying to send a message from a shard to that process to run the code for the restart.

[prisma jewel]

?docs shardingmanager

[young totem]

[prisma jewel]

wait there isn't a shard death capture

When my shards die they respawn on their own, i don't think I understand what you're asking

[young totem]

I already have the code to actually do the restart. The thing I still need is a way to send a signal from a shard to the manager to run that code.

[prisma jewel]

I think if you wanted to restart a specific shard, you could just simply kill the shard

shard.kill

and it will respawn

[young totem]

That's not what I need. Again, I already have the code for the restart. Due to what it does I can't run it on the shard so I'm running it from the parent process. Hence why I need to send a signal to it.

[prisma jewel]

manager to shard or shard to manager

[young totem]

shard to manager

[next plaza]

use client.shard.send()

this will send a message to your sharding manager and you can listen on it with manager.on('message')

[prisma jewel]

Ahhh, I knew I was forgetting something, .send is your best bet

[young totem]

What would the message be? Can I just use whatever? And how would I receive the message?

[prisma jewel]

@next plaza I meant to ask you, if I wanted to index the database on the shard would I just do client.shard.db?

[prisma jewel]

?docs shard#send

[prisma jewel]

There should be documentation from there

[young totem]

Yes I have the docs open now. The type says any which is why I wanted to make sure there aren't any requirements for the format.

[next plaza]

lol what? , no, you do not have to load the database on the client, you can run the database on every shard, btw are you using sql?

[next plaza]

when you use child processes, you can send nearly everything

[prisma jewel]

better-sqlite3, and I think you misunderstood me. I'm asking if I want to pull data from the database that is locally on the shard, would I do client.shard.db

client.db returns null

[next plaza]

on worker_threads you cant bc of the structured cloning algorithm, you have to stringfy them first

[young totem]

Ok. And to receive it I would just listen for a message event?

[next plaza]

yes

[young totem]

Got it, thanks for your help!

[next plaza]

why do you have to pull the data to a another shard?

[prisma jewel]

no no, I have to pull data on the current shard

[next plaza]

it comes up, where you load it, on which variable

[prisma jewel]

its in my client.js under client.db = require('./utils/db.js'), however client.db returns null

[next plaza]

how did you loaded it?

[next plaza]

and what did you executed?

[prisma jewel]

this

[next plaza]

bruh, you didnt exported this file

[prisma jewel]

last 4 lines export settings and users

which are the 2 data tables

[next plaza]

and when you broadcast (this.db), it returns null?

[prisma jewel]

client.db, as its in a client.on('guildMemberAdd', async(client, member) => {})

[next plaza]

how did you bind your events?

[prisma jewel]

filesync

[next plaza]

and why is it client.on ?

[prisma jewel]

what should it be

I'm still migrating to shard code and I'm generally unfamiliar with it all.

loadEvents(path) {
    readdir(path, (err, files) => {
      if (err) this.logger.error(err);
      files = files.filter(f => f.split('.').pop() === 'js');
      if (files.length === 0) return this.logger.warn('No events found');
      this.logger.info(`${files.length} event(s) found...`);
      files.forEach(f => {
        const eventName = f.substring(0, f.indexOf('.'));
        const event = require(resolve(__basedir, join(path, f)));
        super.on(eventName, event.bind(null, this));
        delete require.cache[require.resolve(resolve(__basedir, join(path, f)))]; // Clear cache
        this.logger.info(`Loading event: ${eventName}`);
      });
    });
    return this;
  }

[next plaza]

and you have the event on another folder ?

[prisma jewel]

guildMemberAdd.js

[next plaza]

it should be

module.exports = async(client, member) => {
}

[prisma jewel]

its exporting it all

it all worked before I started sharding, the second shards started, it broke my code partially, all of the breaks happening in the database

[next plaza]

how?

[prisma jewel]

the same way you said

[terse cairn]

hey @crisp coyote, why did you remove <ShardClientUtil>.ids in kurasuta?

[misty cosmos]

var gs = []; client.guilds.cache.forEach(g => {gs.push([g.id, g.memberCount]);}); gs.sort(function(a, b) {return b[1] - a[1];}); var msg = ""; for (var i = 0; i < 11; i++){msg = msg + String(i + 1) + ": " + client.guilds.cache.get(String(gs[i][0])).name + " at " + String(gs[i][1]) + "\n";} message.channel.send(msg);```
i have made this lil code to check the top 10 largest servers my bot is in, but it doesn't work as my bot is being sharded. could someone help?

[dense cape]

I would make use of broadcast eval to get the top 10 guilds from each shard, then pick the guilds with the highest member count of all the shards until you get the true top 10. From there, just map and join the array to however you want it to be

[misty cosmos]

how do i do it, exactly? broadcast eval for the client.guilds.cache.forEach and will it do so?

[dense cape]

You’ll want to return the top 10 guilds from each shard

[crisp coyote]

I didn't remove anything

[terse cairn]

<ShardClientUtil>.ids is undefined

it should be an array

[crisp coyote]

an array of shard IDs?

for that use ShardClientUtil#shards with Kurasuta

i specifically didn't call it IDs cause it would mix Cluster and Shard terms up

which is bad in my opinion

[terse cairn]

oh okie thanks

[sweet lantern]

I'm using client.guilds.cache.forEach to fetch memberCount of each guild and show total users.

Question is, what will i need to do if i'm sharding the bot?

[terse cairn]

@crisp coyote Is there a support server for kurasta?

Because my friend encountered an issue using the clustering package

[next plaza]

👀 , ..., ask you question here , then they can help

[terse cairn]

Ahh okies

[sweet lantern]

Anyone?

[broken sequoia]

https://discordjs.guide/sharding/#broadcasteval this section of the guide shows exactly that

[terse cairn]

Hi

I’ve a question

What’s shards ?

[glossy bluff]

Instances

[sweet lantern]

Okay thanks!

[prisma jewel]

Shards are child processes to shardingManager. Its basically breaking your bot into X amount of smaller bots and logging them into the same client. It keeps the bot working efficiently

[prisma jewel]

So is .send a valid method to use to send data to the shard

or only data from the shard.

[echo fulcrum]

Can ShardingManager actually run typescript file?

it return SyntaxError: Cannot use import statement outside a module when running the typescript file

[next plaza]

technically it should, but you have to use npx and pass it in the shard args

[serene hill]

how do i fix?

CODE

'use strict';

const { token } = require('./config.json');
const { Intents, Discord } = require('discord.js');

const sharder = new Discord.ShardingManager(`${process.cwd()}/shard.js`, { token, respawn: false });

sharder.on('launch', shard => console.log(`launched ${shard.id}`));

sharder.spawn();

ERROR

const sharder = new Discord.ShardingManager(`${process.cwd()}/shard.js`, { token, respawn: false });
                            ^

TypeError: Cannot read property 'ShardingManager' of undefined
    at Object.<anonymous> (c:\Users\prtiu\Desktop\shard-test\sharder.js:6:29)
    at Module._compile (internal/modules/cjs/loader.js:1068:30)
    at Object.Module._extensions..js (internal/modules/cjs/loader.js:1097:10)
    at Module.load (internal/modules/cjs/loader.js:933:32)
    at Function.Module._load (internal/modules/cjs/loader.js:774:14)
    at Function.executeUserEntryPoint [as runMain] (internal/modules/run_main.js:72:12)
    at internal/main/run_main_module.js:17:47

[Done] exited with code=1 in 0.786 seconds

[fathom hamlet]

.intents

[serene hill]

?

[fathom hamlet]

.intents, not .Intents

[serene hill]

ohh

CODE

'use strict';

const { token } = require('./config.json');
const { Intents, Discord } = require('discord.js');

const sharder = new Discord.ShardingManager(`${process.cwd()}/shard.js`, { token, respawn: false });

sharder.on('launch', shard => console.log(`launched ${shard.id}`));

sharder.spawn();

ERROR

const sharder = new Discord.ShardingManager(`${process.cwd()}/shard.js`, { token, respawn: false });
                            ^

TypeError: Cannot read property 'ShardingManager' of undefined
    at Object.<anonymous> (c:\Users\prtiu\Desktop\shard-test\sharder.js:6:29)
    at Module._compile (internal/modules/cjs/loader.js:1068:30)
    at Object.Module._extensions..js (internal/modules/cjs/loader.js:1097:10)
    at Module.load (internal/modules/cjs/loader.js:933:32)
    at Function.Module._load (internal/modules/cjs/loader.js:774:14)
    at Function.executeUserEntryPoint [as runMain] (internal/modules/run_main.js:72:12)
    at internal/main/run_main_module.js:17:47

[Done] exited with code=1 in 0.786 seconds

[fathom hamlet]

idk

[echo fulcrum]

Then is it possible to run with yarn v2?

[next plaza]

const Discord = require('discord.js')
or 
const { Intents, ShardingManager } = require('discord.js');
new ShardingManager....

[next plaza]

try and see lol

[echo fulcrum]

how can i do this ?

[serene hill]

so now i have 0 shards so how can we increase it?

[broken sequoia]

no, and its not related to sharding nor djs

[trail monolith]

Is the shader hard to use?

[next plaza]

no

[next plaza]

in the ShardingManager options or in the manager.spawn funtion

[willow steeple]

Why is the timeout 30s for shard respawning? why doesn't it take the initial spawn timeout value?

[prisma jewel]

Manual shard count

[rough bone]

How can I get the shard id of a client?

Nevermind, just found out the first element in <Client>.shard.ids is the client’s shard id

[wispy ocean]

As an example, if you wanted to split the connection between three shards, you'd use the following values for shard for each connection: [0, 3], [1, 3], and [2, 3]. Note that only the first shard ([0, 3]) would receive DMs.
from the discord dev docs; I'm curious, is it possible to split certain intents for each shard, or is this just something that happens that can't be controlled? (So shard 0 receives x intents, and shard 1 receives y intents and so on)

[spare imp]

@wispy ocean on Discord's end they're all separate, unrelated connections so yeah you can do that

you can't do it with djs internal sharding though, and i assume you can't with shardingmanager either

[wispy ocean]

alright ty

[terse cairn]

How do we get the number of shards from broadcastEval?

[wispy ocean]

You can just get it from <client>.shard: https://discord.js.org/#/docs/main/stable/class/ShardClientUtil?scrollTo=count

[sudden sequoia]

and what does that have to do with sharding ?

[robust mountain]

just awesome help...

[serene hill]

what if we increase shards?

[prisma jewel]

The amount of guilds per shard decreases

[serene hill]

what if there are lot's of shards in 1 guild?

[prisma jewel]

It won’t let you launch more shards then you have guilds

[glacial basalt]

anybody know how to fetch guild number ext. from the sharding manager in the shard file? (I want to make an express server that shows stats about the bot from the main shard file but do not know how to fetch the number for all of the shards)
client.shard.fetchClientValues('guilds.cache.size') only works for individual shard clients

[marsh finch]

You want to use this then: https://discord.js.org/#/docs/main/stable/class/ShardingManager?scrollTo=fetchClientValues

[glacial basalt]

ah okay, tysm

[glacial basalt]

{"status":"200","shards":"2","guilds":"1,3"}```

require('dotenv').config()
const {ShardingManager} = require('discord.js');
const manager = new ShardingManager('./bot.js', {token: process.env.TOKEN})
const shardcount = 2
const express = require('express');
const app = express()

const AutoPoster = require('topgg-autoposter')

const ap = AutoPoster(process.env.TOPGG_TOKEN, manager)

ap.on('posted', () => {
  console.log('Posted the epical stats to Top.gg!')
})

manager.on("shardCreate", shard => {
    console.log(`I've launched the epical shard ${shard.id}.`)
})

manager.spawn(shardcount, 10000);

app.get('/', async (req, res) => { 
    const guilds = await manager.fetchClientValues('guilds.cache.size')
    res.header("Access-Control-Allow-Origin", "*");
    res.json({
        status: '200',
        shards: `${shardcount}`,
        guilds: `${guilds}`,
   })})

  app.listen('6969');```

anybody know how can I get the total number of servers instead of per shard?

[winged bobcat]

Cant, instead reduce and add the values from the fetchClientValues Promise array

[glacial basalt]

okay

[winged bobcat]

manager.fetchClientValues('guilds.cache.size') .then(results => console.log(`${results.reduce((prev, val) => prev + val, 0)} total guilds`)) .catch(console.error);

[glacial basalt]

Ty

[winged bobcat]

Np

[sharp scroll]

enable the intent toggles in the developer portal

[glacial basalt]

can anybody help me?

[dense cape]

Add it up using Array.reduce or a loop

[glacial basalt]

(error)

[glacial basalt]

already did, i'm asking about smthn else 😅 ^

[dense cape]

Code?

[glacial basalt]

client.on('guildCreate', async (client) => {

  const guilds = await client.shard.fetchClientValues('guilds.cache.size')
  const guild = client.guilds.cache.get('822467921653137438')
    const channel = guild.channels.cache.get('845704285656186900')
    channel.setName(`🌍Guilds: ${guilds.reduce((prev, val) => prev + val, 0)}`)
})```

[dense cape]

guildCreate gives you a guild, not a client

[glacial basalt]

oh right, but than how would I do it?

[dense cape]

Prob shouldn’t be updating a channel with live data like that tho, the rate limits are pretty hefty

[glacial basalt]

I know but the bot isn't added to servers often so it shouldn't be an issue

[dense cape]

And when it does?

[glacial basalt]

I guess you're right...

How about making it update in an interval? @dense cape in a ready event (such as every hour ext...)?

[dense cape]

I guess that’s fine

[glacial basalt]

but how would I do it? I plan to take it from my API and update it via a different bot but would that update the channel?

@dense cape

[dense cape]

The interval should be made in the manager file

[glacial basalt]

managers can updated channels?

[dense cape]

No, but you can tell the shard to update the channel

[glacial basalt]

right

[steady nimbus]

There is a way that if the Shard collapses it will reignite

[wet topaz]

it does that by default

[steady nimbus]

@wet topaz so why sometimes the Shard collapses on me and never comes up again

[next plaza]

which version are using of discord.js?

[terse cairn]

how can i access a node module inside of broadcastEval?

[tacit spoke]

Good afternoon, I use the shardList function to spawn shards on different devices, but I can't get the total number of servers in all shards, only those of that device, is there a way to get it?

[terse cairn]

UnhandledPromiseRejectionWarning: #<Response>
pain

[winged bobcat]

Thats not possible since non of the managers talk to each other, you could impliment:

1. Your own ipc for the managers or shards to talk to each other
2. Websocket connections same as above
3. Have the stats stored in a database

[terse cairn]

why am i getting this error

⬤  debug     [WS => Shard 1] Shard was destroyed but no WebSocket connection was present! Reconnecting...
⬤  debug     [WS => Shard 0] Tried to send packet '{"op":3,"d":{"activities":[{"type":3,"name":"t!help | Shard 0"}],"afk":false,"since":null,"status":"online"}}' but no WebSocket is available!
⬤  debug     [WS => Shard 0] [DESTROY]
    Close Code    : 4000
    Reset         : false

[terse cairn]

pls help

[echo fulcrum]

how can i kill all the shards that created from the sharding manager?

[terse cairn]

nvm

[winged bobcat]

You can loop through all the shards and kill them on by one

?docs ShardingManager#shards

[steady nimbus]

12.5.3

[dense cape]

You can’t, that’s not how sharding works

Shards will receive events from certain guilds depending on its id and the number of shards ur using

Prob manually fetch the recommended shards count and divide by 2

One will round down and the other will round up and spawn that many shards

[prisma jewel]

so a larger issue still, I can't seem to pull from my database on shards, I use better-sqlite3

[rain mica]

Hi. I'm sharding my bot with typescript, and get this error

Also I extend Discord.ShardingManager class, if it matters

[static tusk]

XD

[next plaza]

Are u using child processes?

[next plaza]

could you show your shard.js file?

[rain mica]

import Client from "@/structures/Client";

const bot = new Client(process.env.TOKEN, {
    partials: ['MESSAGE', 'CHANNEL', 'REACTION', 'USER', 'GUILD_MEMBER'],
    messageCacheLifetime: 10800,
    messageSweepInterval: 300
})

bot.start()

but shard works normally

[next plaza]

your sharding manager file

[rain mica]

import Manager from "@/structures/Manager";
const manager = new Manager('dist/src/shard.js', {
    totalShards: Number(process.env.TOTAL_SHARDS),
    token: process.env.TOKEN,
    mode: 'process'
})

manager.start()

import Discord from "discord.js";
import ManagerLoader from "@/util/ManagerLoader";


export default class Manager extends Discord.ShardingManager {
    public file: string
    public supportGuildID: string = '712012571666022411'

    constructor(file: string, options?: any) {
        super(file, options);
        this.file = file;
        global.manager = this;
    }

    load() {
        ManagerLoader.loadEvents()
    }

    async start(): Promise<any> {
        this.load()
        await this.spawn(this.totalShards)
    }
}

this is manager class ^^

[terse cairn]

When doing using djs' way to set shards to auto. Do I set auto as string?

[dense cape]

Yes

[terse cairn]

What do I need to change in my codes (command etc.) if I add shard to my bot

[stoic remnant]

there's a guide on it, check the d.js guide

[terse cairn]

Pls send it

[stoic remnant]

https://discordjs.guide, it's a topic in the sidebar

[terse cairn]

Ty

[glacial basalt]

is there any way to disable the autorestart on error with djs sharding manager? A simple API error (SEND_MESSAGES) can crash my bot what is well... bad

[humble shadow]

is there a way i could check if any of my shards are down without the event?

[prisma jewel]

?docs ShardingManager

[sudden sequoia]

there is no built in way do that, you will have to use client/sharding manager options + your own way for communication between these machines/ d.js client/shards

[coral urchin]

can you show the code causing this error?

[terse cairn]

ok wait i fixed the error, but the bot is just now spamming the same sharding error, btw this is the first time for me to try sharding ||although the bot is only on 1 server||

[terse cairn]

@coral urchin can u help me?

[coral urchin]

i actually dont know much about sharding asdlkjgngv

[terse cairn]

lol

i know absolutely nothing about sharding

[terse cairn]

can anyone help me?

[spice frost]

(node:119388) UnhandledPromiseRejectionWarning: Error [SHARDING_READY_DIED]: Shard 1's process exited before its Client became ready.
Any idea?

[terse cairn]

your sharding code?

[spice frost]

const { ShardingManager } = require('discord.js');
const manager = new ShardingManager('./bot.js', { token: [token], totalShards: 2, });

manager.on('shardCreate', shard => console.log(`Launched shard ${shard.id}`));
manager.spawn();```

that's the main file code

[marble maple]

try addind mode to the options

[spice frost]

mode?

[marble maple]

https://discord.js.org/#/docs/main/master/typedef/ShardingManagerOptions

[terse cairn]

maybe to -1

[stray night]

in glitch.com code ${user} this is command of mentioning user then what is the command to mention user's avatar/profile can anyone help with it?

[spice frost]

what

I tried process but still throws the same error, worker throws another canvas error for some reason

[regal oriole]

what am I doing wrong?

🤔

[winged bobcat]

Thats the wrong method

[winged bobcat]

You are trying to get the value of a property on the shard client not execute a function

Use that

[regal oriole]

oh! I understood. XD
Thanks!

[winged bobcat]

Np

[regal oriole]

Oh. I have a problem

I'm trying to get User and send him message. This method returns object without functions.

I... of course, I can use API request or something like but...

[dense cape]

Send inside of the eval

[regal oriole]

genius

omg thanks

[dense cape]

Why are you getting the user via broadcastEval? Is this being called from the sharding manager?

[regal oriole]

i'm tryin' to get user from another shard

[dense cape]

Users aren’t affected by shards

Only guilds and guild* stuff (guild member, guild emojis, etc)

You can just do await client.users.fetch("id") to get the User

[regal oriole]

I have two shards

[dense cape]

Both shards can have the user as well

[regal oriole]

idk. first shard can't

[dense cape]

Could also be none

[regal oriole]

I think that only if the user is on at least one of the shard's servers.

otherwise returns null

ok, maybe im wrong...

[dense cape]

It’s just easier to fetch from client using the method I described earlier

Don’t have to worry abt double dming or not even dming

[regal oriole]

oh. its working. thanks

[regal oriole]

I understood. client.users.cache.get - to get user in client's cache
client.users.fetch - just to get user.

😅 i'm dumb

[dusky osprey]

Can someone link me an extremely detailed resource to what sharding is. I tried googling but didn't understand much

[lusty smelt]

Why are my internal shards dying silently? This is the 4rth downtime in 3 days. One shard just dies silently and the rest follows. According to one of the issues in djs github - i install zlib-sync. Still no luck. Tried listening to SIGABRT - nope.

Currently at 5 shards (internal)

[marble maple]

https://discord.com/developers/docs/topics/gateway#sharding

[next plaza]

downgrade to stable, one of the reasons I did not upgraded to master branch, bc there are some problems with the websocket manager
btw do you get any errors?

[lusty smelt]

Nope none. But i'm using stage channels for the music system & sticker detection, so ig downgrading is not an option

[timber roost]

How can I put the current shard id to the bot's status?

[next plaza]

you just have to add stage channels as structure, you can extend the channel struture and you have extend the message structures for sticker

[lusty smelt]

Perhaps there are files i should replace in the master branch code with the files from the stable branch just for the stability of the branch?

[lusty smelt]

The whole voice system changes, not just the structure. They introduced a base voice channel structure which ig affects several internal files?

[marble maple]

for stickers u have to extend message afaik

[next plaza]

#app-commands message

[timber roost]

Then how can I find the id of the shard the current process is on?

[next plaza]

yes, you have to change a lot, Discord.js has a very nested structure. But thats the only solution...

[next plaza]

listen to the shardReady event client.on('shardReady', shardID)

[lusty smelt]

Rip

[next plaza]

It took me 10 files to add Slash Commands and Interactions like Buttons on v12, and I didnt even fully add every feature, like the CommandApplicationManager

But when you do it, I can help you lol

[lusty smelt]

perhaps moving to traditional sharding will help?

[marble maple]

just asking which mode do u use?

[next plaza]

the problem will be there too, but the percentage, that a Shard silently dies, would decrease. Since there is just 1 Websocket manager for each process (shard)

[lusty smelt]

wdym mode

[marble maple]

i mean process or worker ?

[next plaza]

use process, In woker you will have some problems, bc they have an special cloning algorithmus (which doesnt support objects)

[lusty smelt]

those are internal shards, websocket shards

[marble maple]

hm

[marble maple]

i prefer worker tho : (

[lusty smelt]

Welp, i'll use winmerge to compare the files in the stable and master branch and see if i can bring up a solution since you've said that the stable branch avoids this issue

[next plaza]

I had 10 internal shards running for over 10 days and they didnt had any problems

[lusty smelt]

stable branch, right?

[timber roost]

I'm getting Error [SHARDING_NO_CHILD_EXISTS]: Shard 0 has no active process or worker. when I do ```js
manager.on('shardCreate', (shard) => {
console.log(Launched shard ${shard.id})

shard.eval(`
    this.user?.setActivity({
        type: 'WATCHING',
        name: \`Shard #${shard.id + 1}\`,
    })
`)

});

[next plaza]

you have to first wait until the shard starts

[next plaza]

yes

[timber roost]

How can I do that inside the shardCreate event?

[next plaza]

you can do it on the client event, is much more easier

[timber roost]

But how am I supposed access the shard id that way?

[marble maple]

is the only shardCreate event only?

imean how many events are there for the sharding manager

[next plaza]

listen on client.on('shardReady', (shardid) => { client.user.setActivity( ` Shard ${shardid}`, { shardID: shardid }); })

[timber roost]

I didn't know that existed, thanks!

[lusty smelt]

Well upon comparison, I dont see any major changes that would be the cause of the issue

[next plaza]

you have to compare everything, which is related to the websocket, it mostly crashes bc of an TypeError on connection.url

[lusty smelt]

o

[proud bridge]

You can add stage channels to discord.js v12 with discord-stage npm package

[lusty smelt]

um in which file is it?

[abstract elm]

How would i get information of each shard i.e each shard's ping etc

[radiant sail]

what even is a shard, never heard of them until i read the docs but there was no explanation lmao

[terse cairn]

When you have a lot of guilds (sharding is mandatory at 2500) you can split up your bot, so each shard/instance of your bot handles around 1000-1500 guilds.

[radiant sail]

oh, don't need that my bot is trash lmao

[lucid musk]

Hi, How to use the shardIDForGuildID method

[dense cape]

You give it the guild id and shard count and you get the shard id that will receive events from that guild

[lucid musk]

Oh, thanks

Why does it return an error?
this.client.shard.shardIDForGuildID(${message.guild.id}, ${this.client.shard.count})

[dense cape]

The count should be a number

[lucid musk]

TypeError: this.client.shard.shardIDForGuildID is not a function

[dense cape]

It’s a static method

ShardClientUtil.shardIDForGuildID(…)

[lucid musk]

Thanks

[next plaza]

I will look today at it, I will ping when I find something xD

[lusty smelt]

ty 😄

[quick plinth]

how can I get the ID of a shard from within the shard

[marble maple]

<shard>.id

[quick plinth]

https://i.callumdev.xyz/ln7mh.png

[marble maple]

try ids

[abstract elm]

How would i get information of each shard i.e each shard's ping, etc?

[marble maple]

something like this should help

[abstract elm]

Hmm, Helps me a wee bit but getting undefined:

[marble maple]

ehhh

i think remove ping[9]

[marble maple]

use await ofc

try a async function

[abstract elm]

yep

[marble maple]

mhm now i gtg

[abstract elm]

Odd.

[marble maple]

ofc it’ll be odd u have to call that function

[round vale]

?sharding

[icy crow]

const { ShardingManager } = require('discord.js');
const Manager = new ShardingManager('./bot.js', { token: 'your-token' });
Manager.spawn();

https://discord.js.org/#/docs/main/stable/class/ShardingManager
For an in-depth guide: https://discordjs.guide/sharding/

[abstract elm]

Yeah, forgot about that. Been a hard day today haha. But yeah, nothing.

[marble maple]

const pinged = await client.shard.fetchClientValues('ws.ping')
return pinged

that should work inside the function, if not try it in actual code,
also i have tutions rn : ( sorry

[abstract elm]

Just returning undefined. I'll have a go in the code itself but my hopes aren't too high.

[marble maple]

also inside the eval command put this

let output = eval(text);
if (output instanceof Promise) {
output = await output;
}

this will make sure ull not get promise's

[abstract elm]

Right. Okay.

[abstract elm]

So that's what i added (whether to add the ping() just responded with the promises), without it just returns undefined :/

[marble maple]

aaand u have to call it
ping()

[abstract elm]

Thought i did. Though like I previously said, returns the Promise part of which I'll fix when im next on pc.

[marble maple]

yes then ull to fix the promise part

then itll show results 😄

[abstract elm]

Oh, great.

I'll sort that out, now.

@marble maple You're a star, thank you!

[marble maple]

np : )

[robust zenith]

is there anyway to speed up bot launching with multiple shards?

[lyric lantern]

Earlier I've seen a VSCode plugin here which helps the correct syntax highlight for the escaped broadcastEval snippet, but I can't find it now

Does somebody knows which plugin is that?

[tough idol]

the #1 shard for my bot keeps switching status from idle to ready about every few weeks randomly.
does anyone have an explanation to why this happens?

the bot runs up to 7 shards at a time, so it's weird how it only happens with the second to first shard

[celest grotto]

Hello , i sharding my bot and i dont know how to send a embed in a custom channel :(

[terse cairn]

If the channel is not on the shard, you can use broadcastEval;

...broadcastEval(`this.channels.cache.get(id)?.send({ embed: ${embed.toJSON()} })`)

The ?. will make sure it only executes .send if the channel was found, which should only be the case on 1 shard

[celest grotto]

i use this and give this error

[terse cairn]

Well you shouldn't call it on null

You can only use client.shard if the process was spawned from a ShardingManager

[celest grotto]

nice

[terse cairn]

How did you shard your bot?

[celest grotto]

i make a file

and add this code

const manager = new ShardingManager("./bot.js",{
    token : 'token',
    totalShards : 2

})
manager.spawn()
manager.on('shardCreate', shard => {
    console.log(`Shard #${shard.id} in onine!`)

})```

[icy tree]

If you plan on running the bot on one server do I need to shard my bot?

[dense cape]

No

[restive vortex]

sharding is for bots in 2500+ guilds

[next plaza]

nope, when your bot is over 100k Servers, then you will get Big Bot Sharding, which allows you to spawn 16 Shards in a row instead of 1

[spice frost]

(node:86) UnhandledPromiseRejectionWarning: Error [SHARDING_READY_TIMEOUT]: Shard 0's Client took too long to become ready.
any idea?

[tawny silo]

@spice frost client.on('warn', console.log)
client.on('debug', console.log)

then show me the output

[fringe barn]

can anyone help me with enmap database? it works well but not the bot is running in shards

[spice frost]

but log what

[terse cairn]

just put console.log, that’s it, it’ll log what it needs to

[loud badger]

Hey all, what's the benefit of sharding

[iron sorrel]

None. Sharding is a requirement for bots over 2500 guilds. It doesn't inherently have any benefit.

[loud badger]

Oh I see

pm2 allows for sharding but it would run my bot in parallel, which makes it respond multiple times to a single command

How would I circumvent this if I wanted to support sharding

[marble maple]

use discord js’s sharding

pm2’s shard will be just local, ull have to use discord js or any other package for discord sharding

[unkempt onyx]

I used the sharding from the guide, but im still getting multiple responses for one command, how can i fix this

[tough idol]

did you follow the guide to-point? your code matching the guide with nothing extra?

[unkempt onyx]

Yes

Nothing extra and nothing less, it launches two shard #0 tho i dont know why

[tough idol]

are you using a process manager? like pm2?

[unkempt onyx]

How would i know?

Cause I've never used any package like that

[tough idol]

do you use node in command-line, like node index.js

[unkempt onyx]

Sometimes i get stuck in an endless loop and then the bot keeps responding endlessly still i stop it

Yeah i use the command line

[tough idol]

of course.
check these:

1. accidentally running the process twice?
2. accidentally accessing your bot's entrypoint from shard manager entrypoint?

[unkempt onyx]

How do i check the second

[tough idol]

does the file that contains ShardManager access (require) the file that logs into discord?

[unkempt onyx]

Yes

[tough idol]

that's bad

show me the line where you're requiring it and where you're using it

[unkempt onyx]

But thats what the guide had, unless i need to remove the client.login from my main file

[unkempt onyx]

Ok ill send it in a bit

[tough idol]

you shouldn't be requiring the bot entrypoint if the bot file has an uncaptured Client creation

by "uncaptured", i mean that Client gets created without needing a call

[unkempt onyx]

ok

const { ShardingManager } = require('discord.js');
const { DISCORD_TOKEN } = require(`./config/config.json`);
const manager = new ShardingManager('./bot.js', { token: DISCORD_TOKEN });

manager.on('shardCreate', shard => console.log(`Launched shard ${shard.id}`));
manager.spawn();

@tough idol this is my sharding code

just remove the params for the manager?

[tough idol]

erm, you're not requiring it.

that's correct

[unkempt onyx]

ok thanks alot man

[tough idol]

that should work fine

[unkempt onyx]

this?

[tough idol]

yes

[unkempt onyx]

it launches two shards with id 0

or do i have to wait a while after closing the command line before starting it up again

[tough idol]

that shouldn't affect it

[unkempt onyx]

this happens

[tough idol]

would you like to take this to DMs? i feel like troubleshooting this will clog up the channel

[unkempt onyx]

if you dont mind

[stiff mauve]

@terse cairn

on the client file (the file where your ShardingManager ISN'T located) are you logging in?

[terse cairn]

this is where i make my client:

[stiff mauve]

if u didn't already do it, make sure u put client.login(token);

otherwise that shard won't login (I think, it doesn't work for me unless I do this, at least)

[marble maple]

u have to do super.login(token)
inside the start function

[terse cairn]

Will djs support clustering in the future?

[terse cairn]

why is sharding rising my bot ping soo much?
like my bot ping now is 2000+

[tough idol]

that shouldn't be the fault of djs.
an unoptimised database handler could be the problem

[next plaza]

on the start its high, but it will get less, thats normal

[terse cairn]

when will it get less?

[next plaza]

after some minutes

[dense cape]

Ws ping or api ping?

[terse cairn]

oh ok

[terse cairn]

ws

[next plaza]

really? did you do client.ws.ping?

[terse cairn]

discord api is fine, but time taken to sent msg it high

[dense cape]

There’s some deps you can install to reduce that, at least for production

[terse cairn]

can u pls tell?

[dense cape]

https://discord.js.org/#/docs/main/stable/general/welcome

[next plaza]

z-lib will probably help

[dense cape]

The page lists them all

[terse cairn]

can i install them all?

[dense cape]

Sure

[terse cairn]

i tried rising the total shards, but i get an error every 2 or 3 seconds saying: 429 hit on route /gateway/bot after all this i changed the total shards back to 1, but it still does this error

[dense cape]

?429-gateway

[icy crow]

Warning : this only applies to 429's on gateway, for general 429 errors, refer to ?429

Getting 429 hit on route /gateway/bot when you add client.on('debug', console.log)?
Reasons :
§ You (or someone else) logged in 1000+ times in 24 hours. If you have auto restart on for your code upon crash, you know the culprit.

Are you using shared hosting like Glitch or Repl.it?
§ If yes, your IP has probably been banned by Discord, and theres nothing you can do rly except contact their support, or migrate to a better VPS.
Tip -> Temporary Cloudflare IP bans can last up to 1 hour. In case of IP ban, you can download your bot's code and host it locally.
§ If no situations above suits you, check Discord Status & contact Discord Support.

[dense cape]

@terse cairn

[terse cairn]

oh no my ip has been prolly banned bc i am on repl.it ......

[dense cape]

Yea, that’s the cons of shared hosting (shared hosting = shared ip bans)

[tough idol]

self hosting sucks when your bot hits so many ratelimits

[wooden lance]

hi , there is problem in shard?

[stoic remnant]

No specific one, or perhaps there isn't a problem with sharding, or discord.js' sharding capabilities. Unless you provide the error, maybe we could conclude there actually is one?

[wooden lance]

Error [SHARDING_READY_TIMEOUT]: Shard 0's Client took too long to become ready.

its was work but i do restart for my bot and the bot cant run

[marsh finch]

You might want to try to figure out why your client is taking longer than 30 seconds to ready. (Maybe the debug event could help?)
Or you could increase / disable the timeout.
https://discord.js.org/#/docs/main/stable/class/ShardingManager?scrollTo=spawn

[wooden lance]

the debug its

[WS => Manager] Fetched Gateway Information
    URL: wss://gateway.discord.gg
    Recommended Shards: 20
[WS => Manager] Session Limit Information
    Total: 1000
    Remaining: 0
[WS => Manager] Spawning shards: 0
[WS => Manager] Exceeded identify threshold. Will attempt a connection in 63486898ms

[marsh finch]

Oh, that'd explain it, you somehow managed to identify 1000 times within the last 24 hours.

[wooden lance]

maybe yes the bot was restart but what can i do now?

[marsh finch]

You can force a reset by commenting out this line (or the whole if block)
https://github.com/discordjs/discord.js/blob/12.5.3/src/client/websocket/WebSocketManager.js#L357

You then will receive a system DM from Discord and then have to generate a new token in the developer portal.

[wooden lance]

remove this line?

[marsh finch]

Yes, then you will exceed the limit.

[wooden lance]

thank you so much @marsh finch

[fervent dragon]

Does sharding take up more ram and cpu usage?

[marsh finch]

Not much I can say there, you might also want to log the debug event of the client.

I don't really know, mostly if anything is off looking.
It shouldn't just stop out of nothing.

[lyric lantern]

Why do I got unexpected token error for a \n in broadcasteval?
My snippet:
admin.send("${messageToSend}\n You got this message, because you're the admin of " + guild.name));
If I erase the \n it works good

[marsh finch]

You need to escape the \ there -> \\n

[lyric lantern]

Thank you 🙂

[true swift]

how do you get the number of servers the bot is from the SHARDINGMANAGER?

[tough idol]

<ShardingManager>.fetchClientValues('guilds.cache.size').then(r => r.reduce((a, count) => a + count, 0))

[terse cairn]

I'm not sure if guild would be defined, I think you need this.guild

[tough idol]

no, with fetchClientValues everything isn't within a this keyword and is accessed directly from the client

[terse cairn]

Ahh nice to know

[marble maple]

is there a shard ready event?

or only shardCreate event is present

[tough idol]

there is a shard ready event, that's called <Client>#ready

[proud mortar]

(event) Client#ready
Emitted when the client becomes ready to start working.

[marble maple]

oki

[next plaza]

yes there is for the websocket shardReady

[tough idol]

that's not a thing

[tough idol]

to add on that, <Client>#shardReady is useful if a single client has multiple shards.
if you clients won't ever have multiple shards, then it's pretty much useless

[tough idol]

are you sending your logs to a server channel?

[terse cairn]

no

is different ids everytime

[tough idol]

that happens whenever you try to fetch a non-existing guild

[sudden sequoia]

thats normal? (if your listening for debug events) as type 13 is stage channels, and discord.js v12 dont support them so debug logs them as cannot find the channels type, new version only fix it not crashing

[tough idol]

it is normal

all debug events are normal

the only things you should be watching out for is error events

debug isn't necessary if you're not experiencing a problem already

[abstract elm]

Any idea of how i can get each of my shards to show whether it's online or offline?

[tough idol]

how do you mean? you'd like to report the shards' statuses?

const status = await <Client>.shard.broadcastEval(`
    this.ws.status
`)
status[0] // the status of shard 0

you can also interchange Client with a ShardingManager

but access it at <ShardingManager>.broadcastEval instead

[abstract elm]

Not their statuses, to show whether a shard is online or offline

[tough idol]

await <Client>.shard.broadcastEval('true', ShardToTest).catch(_ => false)

[abstract elm]

Perfect! Thank You Griefs!

@tough idol I don't suppose you know how to resolve the annoying [object Promise]?

[tough idol]

resolve it how? context please

[abstract elm]

What i have is in the following screenshot, although when i go to run the command i get the [object Promise] response.

[tough idol]

it needs to be awaited

broadcastEval returns a promise

[abstract elm]

Cheers.

[paper obsidian]

Hey! Not sure if this is the correct channel, but I have a little question. I am currently trying to get the ShardingManager to work with TypeScript. I am also using the master Branch, however I don't think this has anything to do with my Issue:
When trying to run the bot TS file itself, it works perfectly fine. But when running it through the ShardingManager, I get an error saying that an import Statement cannot be used outside a module. Does anyone know what causes this Issue or does anyone here have any example on how to use ShardingManager with TS?

[tough idol]

you cannot use import if the file is not .ejs or not a declared ESM

you can declare a file as an ESM in package.json

[paper obsidian]

Could you explain further?

[tough idol]

i think you're trying to import a JavaScript file, but JS realising that the imported file is not an ESModule

[paper obsidian]

I am not working with JS files at all. I am using ts-node... (Nodemon actually, but that runs TS Files through ts-node)

[tough idol]

but you're importing discord.js. that's javascript

does it include a stack trace?

[paper obsidian]

Yes, but discord.js has Type Definitions, doesn't it? And importing discord.js itself in my file with the Shard Manager works. But having any import statement in the file of the actual bot/client doesn't.

[tough idol]

what's the filename of your shardingmanager and client?

[paper obsidian]

My ShardingManager is ./launcher.ts and my Client is ./src/bot.ts

[tough idol]

try renaming bot.js to bot.ejs in the compiled version

just to narrow down the issue

[paper obsidian]

Again, there is no compiled version.

[tough idol]

then compile it, tse relative/path/to/folder

[paper obsidian]

(or well there might be, but not as files - I don't know how exactly ts-node works behind the scenes)

(node:21520) Warning: To load an ES module, set "type": "module" in the package.json or use the .mjs extension.
(Use `node --trace-warnings ...` to show where the warning was created)
A:\Development\bot\src\bot.ts:1
import * as dotenv from "dotenv";
^^^^^^

SyntaxError: Cannot use import statement outside a module
    at wrapSafe (internal/modules/cjs/loader.js:979:16)
    at Module._compile (internal/modules/cjs/loader.js:1027:27)
    at Object.Module._extensions..js (internal/modules/cjs/loader.js:1092:10)
    at Module.load (internal/modules/cjs/loader.js:928:32)
    at Function.Module._load (internal/modules/cjs/loader.js:769:14)
    at Function.executeUserEntryPoint [as runMain] (internal/modules/run_main.js:72:12)
    at internal/main/run_main_module.js:17:47

This is the error I recieve by the way. I already tried setting the type to module in the package.json, which didn't fix the issue.

[tough idol]

edit your tsconfig compilerOptions.module and set it to "commonjs" to use CJS modules instead

[paper obsidian]

Already seen that on SO and did exactly that. Doesn't change anything (in regards to my issue).

[hollow badge]

Hey guys,
I'm having trouble requesting emojis across shards. I already got the emoji object via broadcast, but I don't get a GuildEmoji / ReactionEmoji type, which I need to be able to react with it.

Do you have a solution for me?

https://img.namespace.media/images/2021/06/01/WindowsTerminal_X8qHWRZKHL.png
This object should be the right one right?

[tough idol]

did you try renaming the bot.js to bot.ejs?

compile and show the bot.js code

[paper obsidian]

I am having some errors caused my discord.js when trying to manually compile...

[tough idol]

what errors?

[paper obsidian]

(way more stuff, basically pretty much the same)

[tough idol]

ts and djs version?

[paper obsidian]

Don't really know how to find out the TS Version 😅 Unless a new one released within the last 20 minutes, I'm on the newest stable version though.
DJS Version is 13.0.0-dev.7f0d93a2dacf07b39c358b2cf2f5519a757e43b5

[tough idol]

what's the command you ran?

[paper obsidian]

To compile?

[tough idol]

yeah

[paper obsidian]

ts bot.ts from inside the src Folder

[broken sequoia]

66a6a1fcedc8c35b75a89dbb26edc01317bdd7e2 is the latest commit. this hsould be in #djs-help-v14 btw

[paper obsidian]

Well, this issue doesn't seem to have anything to do with any changes in DJS13 though, does it? If you take a look at my original message and the output I sent a bit above, this is an error coming from TypeScript which occured when running through ts-node. The other issues above only occurred after trying to compile the TS files manually for debugging purposes.

[tough idol]

do tsc -v

[paper obsidian]

4.0.3

[tough idol]

djs' typescript is 4.3.2. but i don't think there's any major changes for that to occur, considering it's the same version

[paper obsidian]

Do you by any chance know any open source bots which use Sharding and TS I could just take a look at for reference?

[bright lynx]

You need typescript 4.1+

[paper obsidian]

Alright. That fixes compiling. Not sure yarn installed an outdated version 20 minutes ago. I compiled the client file itself to a JS file now. What exactly should I do with that now?
Edit: Also, that compiled file does not use import Statements, but require, so I assume the Issue lies somewhere when running the thing through ts-node

[hollow badge]

someone got an idea?

[tough idol]

show the code

[paper obsidian]

https://hastebin.com/dotesiyevo.js

[tough idol]

rename bot.js to bot.ejs

ejs? i means mjs

and then run at entrypoint via node

[paper obsidian]

What exactly do you mean by that?

[tough idol]

node shardmanager

[paper obsidian]

So I should compile the file with my Shard Manager to JS as well?

[tough idol]

oh yes, compile the whole thing

[willow steeple]

what does it mean when my shards keep delaying the connection?

the sharding manager has a delay of 5.5 seconds yet when I launch a lot of shards, they all keep requeueing

[paper obsidian]

Then gives me a completely different error of exports not being defined in bot.mjs (exports.__esModule = true;; which is code generated by the TS Compiler)

[tough idol]

yeah that's normal

rename it back and run entrypoint through node