[0.16.18] Repeated and reproducable desync
[0.16.18] Repeated and reproducable desync
Loaded up 0.16.18 and connected to my server. The save was previously on 0.16.16. I just started walking around the map and the game desynced. I tried re-connecting and the game immediately desynced. I tried restarting the server and the problem persisted. I rolled back to the previous save and I was able to connect again however then it desynced again. It seems to keep happening.
Here is the report from the original desync: https://drive.google.com/open?id=1vyxOu ... EXIPaq7LP8
I have more reports if needed.
And I have attached the last map save before upgrading.
I am using vanilla -- no mods were ever installed.
Here is the report from the original desync: https://drive.google.com/open?id=1vyxOu ... EXIPaq7LP8
I have more reports if needed.
And I have attached the last map save before upgrading.
I am using vanilla -- no mods were ever installed.
- Attachments
-
- _autosave8.zip
- (36.33 MiB) Downloaded 391 times
Re: [0.16.18] Repeated and reproducable desync
After you restart the server, what are the steps to reproduce this?
Re: [0.16.18] Repeated and reproducable desync
You just need to run the save and walk around and look at the map for about 20-30 seconds total, you don't have to build or remove anything. The server was running the Linux headless version and I was using the Windows build from Steam. I was the only player on the server.
Re: [0.16.18] Repeated and reproducable desync
Thanks for the report. I'll look into it.
Every one of these save files that are "reproducible" end up being > 30 MB which makes testing them reallllly slooooow.. :\
Every one of these save files that are "reproducible" end up being > 30 MB which makes testing them reallllly slooooow.. :\
If you want to get ahold of me I'm almost always on Discord.
Re: [0.16.18] Repeated and reproducable desync
You have my sympathies, I know how you feel. Let me know if you have any more questions or if there is any other data I can provide.Rseding91 wrote:Thanks for the report. I'll look into it.
Every one of these save files that are "reproducible" end up being > 30 MB which makes testing them reallllly slooooow.. :\
Re: [0.16.18] Repeated and reproducable desync
I'm glad to see you look into it, i've got the same desync problems on 0.16.18 with two desync-reports of 103MB and 104MB eachRseding91 wrote:Thanks for the report. I'll look into it.
Every one of these save files that are "reproducible" end up being > 30 MB which makes testing them reallllly slooooow.. :\
-
- Manual Inserter
- Posts: 2
- Joined: Wed Jan 24, 2018 6:52 pm
- Contact:
Re: [0.16.18] Repeated and reproducable desync
Same here.
Step to reproduce: Login, wait some time, use the map and thats it.
vanilla. Server restarts don't fix that problem.
Step to reproduce: Login, wait some time, use the map and thats it.
vanilla. Server restarts don't fix that problem.
- Attachments
-
- desync-report-2018-01-24_19-44-51.zip
- (112.58 MiB) Downloaded 345 times
-
- Manual Inserter
- Posts: 1
- Joined: Wed Jan 24, 2018 7:08 pm
- Contact:
Re: [0.16.18] Repeated and reproducable desync
i got the same issue on a dedicated server. Everybody disconnects after like max. 1 minute of playing.
Reconnecting, uninstall install did not help.
Pure vanilla, no mods ever installed
and thx btw. awsome game
Reconnecting, uninstall install did not help.
Pure vanilla, no mods ever installed
and thx btw. awsome game
- Attachments
-
- desync-report-2018-01-24_20-07-24.zip
- desync report
- (73.91 MiB) Downloaded 389 times
Re: [0.16.18] Repeated and reproducable desync
We can't reproduce *any* of these desyncs on our end which means we can't fix it as it is.
I need someone to give me a *small* save file - not one that's 30+ MB - that shows the desync with an exact series of steps that reproduces it. "Start the server, join, and wait some time" is not going to cut it with this
My current theory is the linux headless/client is broken and connecting to it with other game versions triggers the desync. So if anyone can confirm that - that would be great. I personally can't because I only have access to a windows machine and it never desyncs for me on any save I test no matter how long I wait.
I need someone to give me a *small* save file - not one that's 30+ MB - that shows the desync with an exact series of steps that reproduces it. "Start the server, join, and wait some time" is not going to cut it with this
My current theory is the linux headless/client is broken and connecting to it with other game versions triggers the desync. So if anyone can confirm that - that would be great. I personally can't because I only have access to a windows machine and it never desyncs for me on any save I test no matter how long I wait.
If you want to get ahold of me I'm almost always on Discord.
Re: [0.16.18] Repeated and reproducable desync
That's the thing, the steps are literally that simple. I'll make a recording of it tonight.Rseding91 wrote:with an exact series of steps that reproduces it. "Start the server, join, and wait some time" is not going to cut it with this .
By "other versions" do you mean other platforms? I can run a small test tonight with a new map on the Linux headless server and see if it happens right away. Alternatively if there is a debug flag I can launch Factorio with on the client or server, I can do that. I can also send you the password to my headless server if you want to connect to it with your debugger attached to your client.Rseding91 wrote:My current theory is the linux headless/client is broken and connecting to it with other game versions triggers the desync. So if anyone can confirm that - that would be great. I personally can't because I only have access to a windows machine and it never desyncs for me on any save I test no matter how long I wait.
Out of fellow software developer curiosity, is your concern with the file size of the save related to how many entities in the map have to stay in sync? If your Windows machine is version 10, Microsoft's Hyper-V is pretty a decent for running VMs if you ever want/need to install a Linux system.
I will run some tests in a couple hours and will get back to you.
Re: [0.16.18] Repeated and reproducable desync
The issue is I can't run the save with the desync-detection flags turned on without 60 seconds of game simulation taking 10+ hours when the save sizes are so large. So "run it for a few minutes and it will maybe desync" becomes "run it for a few days/a week and it will maybe desync" which simply isn't viable.oracleguy wrote:Out of fellow software developer curiosity, is your concern with the file size of the save related to how many entities in the map have to stay in sync? If your Windows machine is version 10, Microsoft's Hyper-V is pretty a decent for running VMs if you ever want/need to install a Linux system.
I will run some tests in a couple hours and will get back to you.
If you want to get ahold of me I'm almost always on Discord.
Re: [0.16.18] Repeated and reproducable desync
Gotcha makes sense. Thanks for indulging me.Rseding91 wrote:The issue is I can't run the save with the desync-detection flags turned on without 60 seconds of game simulation taking 10+ hours when the save sizes are so large. So "run it for a few minutes and it will maybe desync" becomes "run it for a few days/a week and it will maybe desync" which simply isn't viable.
Re: [0.16.18] Repeated and reproducable desync
So I ran a test with a new blank map and couldn't reproduce the problem. I'm guessing it needs more content than a blank map. However I can still easily reproduce it with the save file I sent.
Here is a video of me reproducing it with the save I posted: https://drive.google.com/open?id=1stvN7 ... GKgPb7PsLC
EDIT: I should say that with all my reproductions the time for it to happen is about the same.
Here is a video of me reproducing it with the save I posted: https://drive.google.com/open?id=1stvN7 ... GKgPb7PsLC
EDIT: I should say that with all my reproductions the time for it to happen is about the same.
Re: [0.16.18] Repeated and reproducable desync
The desync involves biters so you'll need at least some of them for it to happen.oracleguy wrote:So I ran a test with a new blank map and couldn't reproduce the problem. I'm guessing it needs more content than a blank map. However I can still easily reproduce it with the save file I sent.
Here is a video of me reproducing it with the save I posted: https://drive.google.com/open?id=1stvN7 ... GKgPb7PsLC
EDIT: I should say that with all my reproductions the time for it to happen is about the same.
If you want to get ahold of me I'm almost always on Discord.
-
- Burner Inserter
- Posts: 8
- Joined: Fri May 26, 2017 4:40 am
- Contact:
Re: [0.16.18] Repeated and reproducable desync
I've been able to make this small map (~5mb), generating a small amount of pollution and with biter evolution set to 1, and have had a couple of desyncs now between linux headless and macos client. Desync is still completely random though, I can't seem to get it to trigger in any reliable manner - one time I was placing concrete, another I was just idle watching an attack, etc.Rseding91 wrote:The desync involves biters so you'll need at least some of them for it to happen.
- Attachments
-
- 45d15394dba3e2ef74d212620b224f3f00a35e22.zip
- (4.64 MiB) Downloaded 879 times
Re: [0.16.18] Repeated and reproducable desync
With 0.16.19 the desync happens instantly when I join the server. Let me know if you want to desync report from 0.16.19.
Re: [0.16.18] Repeated and reproducable desync
Thanks for all the reports, this should be fixed in 0.16.20.