Jump to content

[41.71][BUG] Zombies spawning in the middle of nowhere.


Briao

Recommended Posts

 

 

Hi guys, I have a dedicated server and I ended up noticing an excessive use of RAM I went looking for what was happening and I found several zombies spawned in the middle of nowhere, even deleting them with the debug they continue to spawn an insane amount of zombies does anyone have a solution for this? Delete some file to reset zombie spawn something like that.


image.thumb.png.09bc7720cd5bbd87ee1317d58b24df42.png

Edited by Briao
Link to comment
Share on other sites

  • 3 weeks later...

I have this problem since new MP dropped too, doesn't matter if the server is vanilla or modded, we just call them "the clones" and I even created a discord channel to report them. Before the new tool arrived (the one to remove zeds within a radius) I just blasted them all with a lot of shotguns and ammo (with ofc, admin) but only if there wasn't various chunks affected, because if that's the case, then wiping the zeds files is the way, resetting them all. And even with the new tool to remove them, some will just ignore the "delete command".

I'm hosting on Linux, more specifically, Debian, using LinuxGSM (https://linuxgsm.com/servers/pzserver/) within OMV/Docker (https://github.com/GameServerManagers/LinuxGSM-Docker).
Other stuff: I have another PC where I'm hosting now, but the old one was:


Intel(R) Core(TM) i5-6500 CPU @ 3.20GHz - 32GB of RAM
OMV Version: 5.6.26-1 (Usul)

OS: Debian 5.10.92-1~bpo10+1 (2022-02-03) x86_64 GNU/Linux

Docker: 5:20.10.17~3-0~debian-buster

There was 2 servers with 15GB dedicated for each one, with max 20 players for some time.

The """new""" one is:
Intel(R) Core(TM) i5-7400 CPU @ 3.00GHz - 64GB of RAM.
OMV: 6.0.30-1 (Shaitan)
OS: Debian 5.16.12-1~bpo11+1 (2022-03-08) x86_64 GNU/Linux
Docker: 5:20.10.17~3-0~debian-bullseye


In this one I'm running 5 servers. 4 with 16 players max and 8GB for each one, and other with 32 max and 20GB for it (the first 2 ones from the old server are hosted in this one).

More other stuff:
3 vanilla servers, 2 running since January.

2 modded ones, not gonna bother with those.


As I said before, I'm dealing with this since the new MP dropped, so is not a thing of the world beign old/damaged or something, because I opened a third vanilla just to test this, didn't last a week before the issue was reported. I have looked into the logs but I don't recall seeing something related to the spawn at all. Anyway, if more info is required, I will submit it no problem.
Last thing: I have scheduled automated reboots for the servers every 6H (stopping the server inside the container, then fully resetting the container).

TL:DR: Clones por todos lados

1.png

5.png

Link to comment
Share on other sites

5 hours ago, Beard said:

Could you share the server logs from the vanilla server this is happening on?
A video of it showing how the hordes respawn in front of you would be very useful as well.

 

I leave the logs and a video. They don't spawn in front of me, or at least that is what I belive, since some times there are soooo many of them (and since they roam, you need to find the "patient 0 chunk") that maybe they were already there. If you need more than the debug logs I will upload them, I also have the backup if this ocassion if you need it, I even have no problem in giving you admin access ingame to this server if you want to check it out.
 

 

clones.png

27-06-22_07-01-43_DebugLog-server.txt 26-06-22_23-25-15_DebugLog-server.txt 27-06-22_14-26-09_DebugLog-server.txt

Link to comment
Share on other sites

2 hours ago, XpaDDer said:

 

I leave the logs and a video. They don't spawn in front of me, or at least that is what I belive, since some times there are soooo many of them (and since they roam, you need to find the "patient 0 chunk") that maybe they were already there. If you need more than the debug logs I will upload them, I also have the backup if this ocassion if you need it, I even have no problem in giving you admin access ingame to this server if you want to check it out.
 

 

clones.png

27-06-22_07-01-43_DebugLog-server.txt 834.76 kB · 1 download 26-06-22_23-25-15_DebugLog-server.txt 866.07 kB · 0 downloads 27-06-22_14-26-09_DebugLog-server.txt 127.59 kB · 0 downloads

I'm also seeing the server constantly spamming save/saveall for some reason. I take it an external script is going awry? 

 

But, to me, this looks like the normal amount of zombies you'd expect to see several months in to the game. Did you change any of the sandbox settings to affect peak day, zombie pop .etc? The duplicate zombies might be due to them spawning in the same zone, for example, rather. If there's only 1 or 2 viable outfits, you can get lots of the same zombie spawning. 

 

And there's this, but it only shows up once:

:21
ExceptionLogger.logException> Exception thrown org.sqlite.SQLiteException: [SQLITE_CORRUPT]  The database disk image is malformed (database disk image is malformed) at DB.newSQLException line:941..
Link to comment
Share on other sites

 

8 hours ago, EnigmaGrey said:

I'm also seeing the server constantly spamming save/saveall for some reason. I take it an external script is going awry? 

 

But, to me, this looks like the normal amount of zombies you'd expect to see several months in to the game. Did you change any of the sandbox settings to affect peak day, zombie pop .etc? The duplicate zombies might be due to them spawning in the same zone, for example, rather. If there's only 1 or 2 viable outfits, you can get lots of the same zombie spawning. 

 

And there's this, but it only shows up once:

:21
ExceptionLogger.logException> Exception thrown org.sqlite.SQLiteException: [SQLITE_CORRUPT]  The database disk image is malformed (database disk image is malformed) at DB.newSQLException line:941..

 

The save spam is because some people complained that their stuff is not beign saved when they disconnect, and when I changed the option "SaveWorldEveryMinutes" to 1 the feedback was positive. This is the same with the second server.

The third one, the one that I openend just to check for this bug, everything is default, because I played with the settings of the others and thought about that, so I just said let it be.
The SQL error is from the first server, and I'm sure that it happened becase the lights went out like crazy for some time (and I really need an UPS), but something needs to be clear about this: when I launched the third server the lights didn't went out for almost a month, and as I said, it didn't last a week before someone reported this stuff, it was a clean one.

 

Maybe some chunks won't reset a timer or something? The only thing that they have in common is that they are clones, won't matter if the zones are visited enough or abandoned for a long time, it can be the next house to yours, some random street anywhere or even your house (yes, it happened to some people). Right now I don't have a report that bad, mostly some lame tiny horde clones, like 20/50 scattered around some houses.

 

This screenshot is from a house close to the base of someone in the third server, he noticed that the fps started to drop like crazy when he was getting close to this house and then just crashes, and when I tried to get there, the game just freezes and I have to force close it. I lowered the graphics, hide the UI and make the window really tiny. They ignore the horde manager stuff (you can see the arrow, indicating that I already tried that) when they are a bunch, but at least the bodies do disappear. I even have a video about it, check it out and tell me if you still think this is the normal amount of zeds or even the place where they spawn is normal. They were soooo many, that the stairs leaked zeds everywhere lol
 

 

3.png

Link to comment
Share on other sites

I don't want to be mean here, but saving every minute means the server is going to pause both the clients and the server every minute. I can't see that being good and likely having unintended consequences.

 

If the problem is that zombies basically "respawned everywhere" after something like a power outage happened, then it's more likely down to file corruption of some kind. It'd be a considerable number given you're far enough into the save to have snow checks firing in some logs, but this is just a guess. How much would be determined by the date and zombie settings in sandbox.

 

Zombies lagging a house is normal. This is why the server culls them down to 500 on one screen when they're 20-odd tiles behind the player and out of sight. As some people tried to defeat this for "immersion," I need to point out that if you happened to disable this, it could be the cause. Otherwise, it just seems like a lot of zombies in a house and that can happen naturally while playing the game, unfortunately.  I highly recommend setting zombie pop to normal or lower than normal values regardless to smooth things out.

 

If none of this seems helpful, then could you try creating a new, vanilla server with all default sandbox settings and letting it run for a while? It might serve as a good control.

Link to comment
Share on other sites

3 hours ago, EnigmaGrey said:

I don't want to be mean here, but saving every minute means the server is going to pause both the clients and the server every minute. I can't see that being good and likely having unintended consequences.

 

If the problem is that zombies basically "respawned everywhere" after something like a power outage happened, then it's more likely down to file corruption of some kind. It'd be a considerable number given you're far enough into the save to have snow checks firing in some logs, but this is just a guess. How much would be determined by the date and zombie settings in sandbox.

 

Zombies lagging a house is normal. This is why the server culls them down to 500 on one screen when they're 20-odd tiles behind the player and out of sight. As some people tried to defeat this for "immersion," I need to point out that if you happened to disable this, it could be the cause. Otherwise, it just seems like a lot of zombies in a house and that can happen naturally while playing the game, unfortunately.  I highly recommend setting zombie pop to normal or lower than normal values regardless to smooth things out.

 

If none of this seems helpful, then could you try creating a new, vanilla server with all default sandbox settings and letting it run for a while? It might serve as a good control.

 

Did you read my post? I said that the third server, the one from where the clip is from, EVERYTHING is default, everything, even that option, and power outage wasn't an issue when it happened on this one. Also we don't see pauses when the server saves, unlike doing /save manually. Lagging a house is normal? Sure, but did you watched the video or saw the screenshot? Are you telling me that having more than 100 zeds under a stair is normal? Where they can't go anywhere? Where they are so much that the collisions between them makes a bunch of others clip through the walls? Even in your own house? The one where you are doing stuff everyday? You clean some place, and tomorrow that place crashes your game because for some reason is filled but now you can't even count them, normal? Every setting by default.


No no no no, don't think that this is normal. I looked for some time the posts in this forums to check if someone was having this problem too, and since there wasn't any, I asume that it has to do with the way I host them, maybe Docker or LinuxGSM are the issue here, and the people playing told me that they didn't see this happening on other servers, at least for the time they were playing, really hard to reproduce so I understand that you would think that this is something X user messed up.
 

Now I'm waiting for a "must wipe" update, since people didn't wanted to lost their stuff (running since January), to try hosting with Windows. For the time being, I will upload more stuff when happens this bad again and hope that OP updates this post with more info, maybe we find some similarities.

Link to comment
Share on other sites

Obviously I read it and found much of it didn’t make sense, so asked for clarification and more details because you threw a lot of varied stuff out there and are even doing something really weird w/ the save system. That’s just not how it’s supposed to be used and I never would have thought someone would do that.

 

But in short, no: I’ve not seen or heard of these problems before and neither has anyone else with a server that I asked. It doesn’t seem related to the op beyond some zombies look similar.

 

I assume its either docker, as you say, or the save thing interfering. That’s really all I can suggest.

Link to comment
Share on other sites

41 minutes ago, EnigmaGrey said:

Obviously I read it and found much of it didn’t make sense, so asked for clarification and more details because you threw a lot of varied stuff out there and are even doing something really weird w/ the save system. That’s just not how it’s supposed to be used and I never would have thought someone would do that.

 

But in short, no: I’ve not seen or heard of these problems before and neither has anyone else with a server that I asked. It doesn’t seem related to the op beyond some zombies look similar.

 

I assume its either docker, as you say, or the save thing interfering. That’s really all I can suggest.

 

Third server had every setting default, and I mean it, EVERY OPTION. I tried to give all the maybe required info, hoping that someone shares the same issue, but yeah, right now there isn't something to work on as it is maybe an isolated problem with my setup.
As for the "SaveWorldEveryMinutes", the tooltip says "Loaded parts of the map are saved after this set number of real-world minutes have passed. (The map is usually saved only after clients leave a loaded area)", it does not indicates any recommended value, common sense told me that sure it will put more stress on the nvme drive, cpu and ram usage causing some problems, but when the third one started doing this with the option untouched I just blew them away. I monitored the cpu and ram usage, they weren't that crazy to justify changing it, rising to max 80% with all the servers running and ram having always like 5~ GB free (not cached, free), accepting that the storage will need to be changed sooner.

 

Anyway, I'm pleased with having this "documented" in the forums.

Link to comment
Share on other sites

I had to use "SaveWorldEveryMinutes" set at 15 on a G-Portal server because it never saved on a server restart. Instead what happened was areas previously cleared would become uncleared and safehouses never saved when I used them to test also. I've seen this problem being reported quite a while back too, nothing seems to come of it. Happened on vanilla server so no, mods can't be blamed.

 

I read it was something to do with how the server was shutdown and restarted. Only doing it a certain way would it actually save, otherwise you get issues like this and being forced to use the "SaveWorldEveryMinutes" to get it to actually save. I've also read you can save with an admin character using the /save command but as admin characters aren't really designed to be used to play the game with, this is an awkward solution.

 

I would rather not have to use "SaveWorldEveryMinutes" because it did lag the server. Hopefully this gets addressed.

 

 

Edited by gromit
Link to comment
Share on other sites

16 minutes ago, gromit said:

I had to use "SaveWorldEveryMinutes" set at 15 on a G-Portal server because it never saved on a server restart. Instead what happened was areas previously cleared would become uncleared and safehouses never saved when I used them to test also. I've seen this problem being reported quite a while back too, nothing seems to come of it. Happened on vanilla server so no, mods can't be blamed.

 

I read it was something to do with how the server was shutdown and restarted. Only doing it a certain way would it actually save, otherwise you get issues like this and being forced to use the "SaveWorldEveryMinutes" to get it to actually save. I've also read you can save with an admin character using the /save command but as admin characters aren't really designed to be used to play the game with, this is an awkward solution.

 

I would rather not have to use "SaveWorldEveryMinutes" because it did lag the server. Hopefully this gets addressed.

 

 


That’s interesting. I’ve never personally seen it reported, so I think this is the first we’re hearing of it, unfortunately.

 

How do you normally shit down the server? It should always be done with /quit at a minimum (save followed by quit preferred).

Link to comment
Share on other sites

50 minutes ago, EnigmaGrey said:


That’s interesting. I’ve never personally seen it reported, so I think this is the first we’re hearing of it, unfortunately.

 

How do you normally shit down the server? It should always be done with /quit at a minimum (save followed by quit preferred).

 

Glad I'm not the only who makes that typo!

 

My old Nitrado server never did this and I never needed to have a value set for "SaveWorldEveryMinutes".

 

G-Portal's back end doesn't actually show what command they're using when you choose to shut down or restart the server. I'm also not the most knowledgeable on this topic, but I'm wondering if they're using ctrl-c as suggested here, which is causing the issue of it not saving?

I'm not sure if that's of any use to you.

Link to comment
Share on other sites

Lol. Autocorrect working it’s magic.

 

I’d assume that yes, it just kills the process from what you described. Most of these companies seem to be almost 1-man bands despite outward appearances, so writing a custom power shell script (or equivalent) to send keyboard input to the console is probably not universal.

 

 Pressing the off button in the web ui might even restart the vm, for all I know. :p

 

ctrl+c just shuts it down immediately when sent. That means the internal save command never runs. Pure luck it doesn’t corrupt the files themselves …

 

… and maybe we just explained how people keep nuking their map_meta.bins causing random house/survivor house rerolls, which can eat bases, or wreck safehouses … ooph. 

Link to comment
Share on other sites

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now
×
×
  • Create New...