Windows Server Stuttering and Low FPS visible in server console (Day 69, 6k Map)

T

Thedarklurch

Refugee
Hi everyone, I'm reaching out here because I'm at a loss. We're on day 69 of the game and are currently having major issues. We keep experiencing micro-stuttering; the zombies don't move smoothly, and when we drive through the world, we repeatedly encounter lag that causes one of us to teleport around.

We have a dedicated server with a hosting provider.

It has the following specs:
Hardware Specs:

  • CPU: AMD EPYC 7443P (4 vCores)
  • RAM: 10 GB (currently using ~6.1 GB RSS)
  • DISK Space: 17.7 / 20 GB
  • Map: 6k World (Day 69)
The Problem (Evidence from Logs):

  • Server FPS: Drops to 15 - 20 FPS during movement (visible in server console [Pterodactyl]), causing AI lag and rubberbanding. [Ingame we have 70 - 80 FPS]
  • Heap Limit: The internal Unity Heap is capped at 3.5 GB (Max: 3530.3 MB), which is nearly full.
  • Warning: WRN [Steam] Tick took exceptionally long: 44 ms.

While we're playing, the server shows the following usage on the dashboard:

CPU Usage: 65.54% / 400%
Memory Usage: 5.97 GiB / 10 GiB
Disk Usage: 17.49 GiB / 20 GiB

We're currently playing on version V2.6 (b14)

and have kept the default settings:

MaxZombies: 64
MaxAnimals: 50
ViewDistance: 12



Now my question is: is the server too underpowered for the game with these settings, or should we upgrade something? We’d like to finally have a smooth gaming experience again. Up until day 30, we hardly had any issues, but starting on day 35, we started having more problems. I’d appreciate any tips that could help me solve this problem.
 
The V2.6 server has a memory leak. We are waiting for TFP to find out the cause and fix it. We have conducted no less than three tests with the original version without any modules and with modules.
 
Blue mame said:
The V2.6 server has a memory leak. We are waiting for TFP to find out the cause and fix it. We have conducted no less than three tests with the original version without any modules and with modules.
Click to expand...
Can you link to an announcement regarding this? I recently had v2.6 eating 64GiB of RAM even when playing as host. The "fix" for me was deleting the two cache folders hidden in AppData/Roaming/7DaystoDie as well as the datastorage folder. I have seen no admission of a confirmed leak, but that would make far more sense.
 
The_Great_Sephiroth said:
Can you link to an announcement regarding this? I recently had v2.6 eating 64GiB of RAM even when playing as host. The "fix" for me was deleting the two cache folders hidden in AppData/Roaming/7DaystoDie as well as the datastorage folder. I have seen no admission of a confirmed leak, but that would make far more sense.
Click to expand...
We were able to discover this because I am a server community administrator myself. I discovered it when I participated in the V2.6 testing during the later stages of the V2.5 stable version. After the official release, all server community administrators in our country discovered the problem, but so far no server community administrator has submitted a report. I submitted a bug report myself. Hopefully, TFP can resolve this as soon as possible. It seems that all platforms' V2.6 servers have map and entity leaks. I don't know what critical logic code they changed in the V2.6 update.
Post automatically merged:

The_Great_Sephiroth said:
Can you link to an announcement regarding this? I recently had v2.6 eating 64GiB of RAM even when playing as host. The "fix" for me was deleting the two cache folders hidden in AppData/Roaming/7DaystoDie as well as the datastorage folder. I have seen no admission of a confirmed leak, but that would make far more sense.
Click to expand...
There was a map memory leak issue in a previous version, I think it was A17 or A18, I forget exactly. They quickly released a patch to fix it. I wasn't very active in this forum at the time, so I don't know if anyone submitted a bug report. However, letting TFP know about this bug should get them to fix it quickly, as it's a critical bug.
 
Blue mame said:
The V2.6 server has a memory leak. We are waiting for TFP to find out the cause and fix it. We have conducted no less than three tests with the original version without any modules and with modules.
Click to expand...
And could that be what's causing the lags? We usually use 6 GB out of 10 GB. Yesterday we were at 8 GB, but we were moving around the map a lot and exploring new areas.
 
Thedarklurch said:
And could that be what's causing the lags? We usually use 6 GB out of 10 GB. Yesterday we were at 8 GB, but we were moving around the map a lot and exploring new areas.
Click to expand...
Once Blood Moon triggers, if more than 5 players are exploring in-game, the server will run out of memory within two to three hours. When the online player count reaches 8, even allocating 20GB of RAM is far from sufficient. As the number of players rises, memory usage grows **exponentially**.I have found a reasonable explanation for why almost no one reports this memory leak bug to TFP. Forum moderators keep insisting that 7 Days to Die officially supports a maximum of only 8 players, largely to avoid accountability for underperforming servers failing to deliver the expected player scale.This rigid narrative has led nearly all server administrators to form a consensus: TFP simply does not care about dedicated server optimization. Consequently, no one bothers to report this critical memory leak issue to the official development team.I witnessed a real-world case on a Windows server: with 15 players online simultaneously, server memory skyrocketed from 5GB to over 40GB in just half an hour.This is exactly what I have consistently emphasized in my previous posts: community moderators should never make absolute, conclusive statements in their replies. Such rigid assertions directly block valid bug feedback from reaching the developers, leaving critical underlying issues unreported and unfixed indefinitely.
 
Blue mame said:
In fact, this is a protection mechanism to prevent low-performance servers from failing to achieve the performance described.
Click to expand...
Given that you have to mod the game to even allow more than 8 players (as far as I understand it; I don't play and have zero interest in multiplayer.)...it's just what they officially support. Yes, it might be for performance reasons, but that doesn't change the simple fact.
 
Vaeliorin said:
Given that you have to mod the game to even allow more than 8 players (as far as I understand it; I don't play and have zero interest in multiplayer.)...it's just what they officially support. Yes, it might be for performance reasons, but that doesn't change the simple fact.
Click to expand...
I'm not condemning anyone, but rather pointing out that if all server administrators accept this rule, any subsequent problems will likely be blamed on my failure to comply with TFP standards for servers with 8 or fewer players. Most people would probably see it that way. The remaining administrators, after understanding the situation, will be conflicted about whether TFP prioritizes server vulnerabilities or performance, and whether my report will cause unnecessary controversy.
 
Blue mame said:
Once Blood Moon triggers, if more than 5 players are exploring in-game, the server will run out of memory within two to three hours. When the online player count reaches 8, even allocating 20GB of RAM is far from sufficient. As the number of players rises, memory usage grows **exponentially**.I have found a reasonable explanation for why almost no one reports this memory leak bug to TFP. Forum moderators keep insisting that 7 Days to Die officially supports a maximum of only 8 players, largely to avoid accountability for underperforming servers failing to deliver the expected player scale.This rigid narrative has led nearly all server administrators to form a consensus: TFP simply does not care about dedicated server optimization. Consequently, no one bothers to report this critical memory leak issue to the official development team.I witnessed a real-world case on a Windows server: with 15 players online simultaneously, server memory skyrocketed from 5GB to over 40GB in just half an hour.This is exactly what I have consistently emphasized in my previous posts: community moderators should never make absolute, conclusive statements in their replies. Such rigid assertions directly block valid bug feedback from reaching the developers, leaving critical underlying issues unreported and unfixed indefinitely.
Click to expand...
And I only have one project with a total of two players and the default map. I don't even want to think about what would have happened if we'd used a bigger or different map. I think it's a shame; I hope a mod reads this and passes the problem along.
 
Blue mame said:
Once Blood Moon triggers, if more than 5 players are exploring in-game, the server will run out of memory within two to three hours. When the online player count reaches 8, even allocating 20GB of RAM is far from sufficient. As the number of players rises, memory usage grows **exponentially**.I have found a reasonable explanation for why almost no one reports this memory leak bug to TFP. Forum moderators keep insisting that 7 Days to Die officially supports a maximum of only 8 players, largely to avoid accountability for underperforming servers failing to deliver the expected player scale.This rigid narrative has led nearly all server administrators to form a consensus: TFP simply does not care about dedicated server optimization. Consequently, no one bothers to report this critical memory leak issue to the official development team.I witnessed a real-world case on a Windows server: with 15 players online simultaneously, server memory skyrocketed from 5GB to over 40GB in just half an hour.This is exactly what I have consistently emphasized in my previous posts: community moderators should never make absolute, conclusive statements in their replies. Such rigid assertions directly block valid bug feedback from reaching the developers, leaving critical underlying issues unreported and unfixed indefinitely.
Click to expand...
This is false, or it was last night. We're running longer days and our first blood moon featured all the glowy zeds (green, blue, gold) and we literally fought them, non-stop for an hour. Only three of us but I never crossed 9GiB of RAM usage. In fact, since clearing the game caches, it has run solid on 2.6.

I saw your bug report but no logs and if you're running mods it isn't helpful, especially that home-brew anti-cheat. We already have VAC and EAC (optional), so a third one in the mix seems crazy to me, but hey, your server. I also know of people who can bypoass a stack of those things and cheat all day, so I prefer the only functional anti-cheat: An actual moderator.

We just played for over two hours, almost three, non-stop, with three of us. We cross all but the snow biome during our session and again, less then 9GiB used.
 
You must log in or register to reply here.
Back
Top