Server Named Muddy Waters Crashing every 20-40 minutes


#1

I’ve been an admin on Muddy Waters, an expert mode survival server, for more than a year now and we’ve been having this problem for the last month. Every 20-40 minutes or so (sometimes even more often), the server crashes and everyone is kicked off (telling everyone they were “disconnected”). Sometimes when a player is in the middle of picking something up, putting something down, or use an item for crafting, the item will dupe or the player will lose the item. No one currently on the server has ever seen the owner log in (even every day players for more than a year) so we can’t ask them for help.

The only thing we have found that gives us a clue is when we run the /debug-log command. We tend to crash when the MB share reaches 930-1300MB. We have tried to reduce the amount running on the server simultaneously and it doesn’t do anything.

Some information that might head off your guesses: We all use different devices. We are from all parts of the world and it doesn’t matter. It can happen when a lot of people are on OR if only one person is on alone. We have tried to resolve the issue by limiting the number of portals, reducing the number of animals (tamed and untamed), resolving glitching spaces we found, and replanting protection signs.

We love our server and have invested a lot just surviving in pure expert mode and to keep it hack free. I’d really appreciate it if we could get some help.

Thank you!


#2

Typically hit or miss.


#3

Happens on @UNIHORSE’s server and other servers too…


#4

Has anything changed in the server in the last month? Most of the servers that are experiencing the disconnects are very developed servers and have been experiencing the disconnects since the update. Has a more of the server been explored? Was flight just achieved? Has building increased or use of a certain building material? Maybe you can provide clues to solve this mystery.


#5

I know this is kinda irrelevant but i cant get to muddy waters because my phone got reset for some reason. (can you please tell to anyone who wants to see me?)


#6

Yes I suppose it has, about 2 months ago more people have gone to visit the poles whereas before I think only the owner had circumvented the world. Now I think 3 people have been traveling around the world. It’s a x16 world so it’s pretty big and I wouldn’t call it anywhere close to “developed”. Just old.

Hey Kevasa- it’s Kunoichi, what would you like us to do?


#7

Edited title to be a little more informative


#8

That’s a good datapoint. I wonder if it’s a detail that Dave could look at.

I feel your pain @vbabe1. I have three old servers that suffer from this same thing. To date, Dave has not discovered the root cause. I don’t know if @Milla has any more info or not. Even as an owner there’s nothing I can do for my servers.


#9

Nothing solid enough to pull him away from Sapiens.


#10

(Tap to enlarge images)

@Starlight1

This is the log I created of the MB Share from the Starlight Skys 4 cloud server.
I started logging the MB Share from 07:00 UTC, and stopped at 07:40 UTC. I tried to get a reading almost every minute. The cells with “0” in them represent when the server crashed.

Both crashes I logged where also within this threshold.


#11

So will the problem ever get fixed?


#12

It will be fixed when the cause for the disconnects is identified.


#13

This is great! Should I do the same? What makes the MB shared go up?


#14

If you want to do the same. I’m sure you will see a similar trend.


If Dave never looks into it the problem probably never will be identified.


#15

I think more data would definitely help. I might try to do that as well. The MB increases as the server has to load and cache more data, from people being online and loading chunks of the map, or inspecting containers, trade portals, shops, etc. Everything the players do will make the server load data.

@Milla, would it take long for Dave to check if there’s some sort of limit with the cloud instances that makes them restart when they hit it? I would think they should be simply starting garbage collection and offload memory that isn’t needed. Yet they seem to increase to a point and then restart. I don’t want to say “memory leak” but there I just said it. :wink: A proper garbage collection management scheme would show more up and down with the memory.


#16

Please Dave, we beg of you! :cry:


#17

I feel you guys and I really feel sorry for you but you better stop waiting and hoping, you better quit this game, they will never fix this crashing problem!


#18

I’m not giving up. The crashes are tolerable to me. I was just messing around as I gathered the stats on Rabbithole. I included the total memory MB which was probably a mistake. But it’s interesting that it’s pretty stable across the top, while the shared and resident memory climbe to a peak and then drop when the server restarts. And the gold line is the loaded macro blocks which fluctuates but doesn’t climb in any way related to the shared memory. For the few peaks that are under 1000 MB, I believe we had a few cases where the last player simply logged off before the server hit its crashing point.


#19

If I quit the game I’ll be bored eventually he will hopefully fix it but it doesn’t affect most of the servers I personally play on a lot some it does


#20

I find it interesting that the server crashes when no one else is around. I had assumed when everyone was off it would fall into some kind of holding pattern.