[Bug] Overflowing ped pool result in a crash

[d22tny]

Summary

I tested the old problem where if you would be overflowing the pools of networked peds / objects / vehicles msgPackedClones would grow exponentially, packet loss will rise and everyone would end up timeouted.

That seems to be fixed, with objects it stops networked creation, they still get created locally, no growing happens and you get No free spaces in object pool. Cannot create new Object network object. -> No problems here.

For vehicles same thing happes except there they won't spawn locally, and no console log is received as it happens in object. -> No problem here.

However, when i did the same test with networked peds, everytime the game froze and no crash was present, just the game closing.

Issue type

Client

Repro rate

Always

Server build version

90

OS

Windows

CPU

Intel i9-12900K

GPU

Nvidia GeForce RTX 3080 TI

RAM

32GB DDR5

Storage type

SSD NVMe

Connection type

Ethernet

ISP and bandwidth

1GBPs

DxDiag

Network graph

Platform

Linux

OS version / distribution

Ubuntu 24.04.4

CPU

AMD EPYC 4584PX

RAM

64 GB

Using txAdmin?

No

Hosting provider

Other GSP

Machine type

Dedicated

/perf endpoint output

No response

DDoS protection

ulimit -n value (Linux only)

No response

Docker Compose file (Docker only)

No response

Steps to Reproduce

1. Make a repro resource to spawn a networked ped every 0.1 s
2. Wait for pool to be full.
3. Spike -> Game closes.

Expected Behavior

Same behaviour as in objects and vehicles.

Actual Behavior

Froze and client exit.

Evidence

No response

Additional Context

No response

[wolfram-cfx]

Thanks, do you have an sentry id when you crash?

[d22tny]

Hi, nope. The game would freeze and then quit, but i will try it again on the latest server build.

[d22tny]

Okay, i did the tests again and i got some more details about the repro.

How i tested this time :

2 players spawning peds, using this code :

local spawnedPeds       = {}
local pedRunning        = false
local PED_MODEL     = `a_m_y_hipster_01`

local function loadModel(model)
    if not IsModelInCdimage(model) then return false end
    RequestModel(model)
    local timeout = GetGameTimer() + 5000
    while not HasModelLoaded(model) and GetGameTimer() < timeout do
        Wait(0)
    end
    return HasModelLoaded(model)
end

RegisterCommand('startpeds', function()
    if pedRunning then return end
    pedRunning = true
    CreateThread(function()
        if not loadModel(PED_MODEL) then pedRunning = false return end
        while pedRunning do
            local p = PlayerPedId()
            local c = GetEntityCoords(p)
            local ped = CreatePed(4, PED_MODEL, c.x, c.y, c.z, GetEntityHeading(p), true, true)
            spawnedPeds[#spawnedPeds + 1] = ped
            Wait(200)
        end
        SetModelAsNoLongerNeeded(PED_MODEL)
    end)
end, false)

If we both stay in the same location, ped creation will stop after reaching the limit, no crash still. If i get out of that zone, it starts creating peds again, if i come towards the first zone of spawning his game freezes, and then exits, after i get closer, same happens to me.

[edwoardtw]

Thanks, do you have an sentry id when you crash?

Still the same issue on recent build 115

[chromium-cfx]

Thanks, do you have an sentry id when you crash?

Still the same issue on recent build 115

Do you get a crash with a sentry id? I'm unable to reproduce this locally.

[wolfram-cfx]

Hi, this is not reproducible anymore after our pool increments. Feel free to reopen if you still encounter it.