Empty ttp files

Z

Zæffy

Refugee
Hello,

I'm am one of the managers of a private server hosted by a server rental company. Since the beginning of the Alpha 20 (including the pre-stable release), we are facing a problem : some player's ttp file turn either empty, or to a 1kb file, and the player are suddenly unable to login.

As far as I can tell, they didn't do anything suspicious. They are usually playing, then start to have trouble / crashes and when they try to reconnect, they cannot.

The server have 0 mods installed, Twitch integration is desactivated, and we created a new map for the A20. We though that it may be an hardware problem, so we asked and got our server moved to another physical machine, but the problem keep being there.

Would you have any idea of why that could happen ?

Thank you !

 
Hello,

Sorry for the delayed answer, I had to wait for the problem to happen again in order to gather all the needed informations.

Here are the informations you asked for :

- Did you start a new game and re-create the issue : yes, we wiped the server a few weeks ago and the issue is still there

- Have you validated your files through Steam : yes, the users who have the problem did reinstall the game, and so did we on the server.

- After validation, did you re-try with a  brand new game : yes (when we reinstalled we wiped)

- What happened / is the problem: sometimes, when players (mostly one since we changed server) try to reconnect, the .ttp file of their profile turn almost empty (here is an exemple: http://gofile.me/5ijmE/Dcb8xiobm ). When they have this problem, they cannot login anymore, and they have this console message when trying: http://gofile.me/5ijmE/xpJVgKvi0 . The only solution we have so far is to remove the ttp file (or replace it with an older save).

- Game type : MP dedi, hosted by g-portal

- There is no mod on the server

Client log: the player that have this issue now have no log (the game doesn't seems to generate them), and I don't know why ?

Server log: https://justpaste.it/7mph4

 
Zæffy said:
No, they don't 😕 . To be more precise, they do have a logfile up until the day they switch from unstable to stable (the day A20 went stable), and none since then.
Click to expand...
Have them do a clean uninstall/reinstall as something is definitely not right there. The procedure for doing a clean uninstall/reinstall can be found by clicking the blue banner at the top of the page. They should also delete the Mod folder manually if present. This will assure just A20 files from stable are added and no files from previous A20 experimental are present. The log file is ALWAYS generated unless stated in the command line as -no logfile (I think that's the command) so if there is no log file being generated, the installation is corrupted needs to fixed by the above clean uninstall/reinstall.

 
Zæffy said:
No, they don't 😕 . To be more precise, they do have a logfile up until the day they switch from unstable to stable (the day A20 went stable), and none since then.
Click to expand...




***************************************
1. what exact pre-stable build did work before you went to the stable build? current stable is a20b238

2. is the server (rented or player based dedi)?

3. is the server set up to use EOS and are the players platform.cfg set to use EOS? (could be mis-match) but still should be able to connect.

4. have players go to their game install directory and click on the 7dLauncher.exe and start game and attempt to join the server.

5. there will be an output log created in ...\7 Days To Die\7 Days To Die_Data folder (find the newest log and paste its content into pastebin (that other site bunches everything up and is an eye sore))

Thanks  :)
QA Tester-unholyjoe
***************************************

 
Last edited by a moderator:
unholyjoe said:
***************************************
1. what exact pre-stable build did work before you went to the stable build? current stable is a20b238

2. is the server (rented or player based dedi)?

3. is the server set up to use EOS and are the players platform.cfg set to use EOS? (could be mis-match) but still should be able to connect.

4. have players go to their game install directory and click on the 7dLauncher.exe and start game and attempt to join the server.

5. there will be an output log created in ...\7 Days To Die\7 Days To Die_Data folder (find the newest log and paste its content into pastebin (that other site bunches everything up and is an eye sore))

Thanks  :)
QA Tester-unholyjoe
***************************************
Click to expand...


Note for clarity: I realise that I may have done a slight mistake while translating in english ; since our last wipe, there is only one player who is still facing the empty .ttp file problem. In addition, every other players have output_log file (e.g. I do have them), it's only this specific player who doesn't.

1/ We always used the latest build, on the unstable branch when A20 was only on unstable, then on stable. The current problem (when the file went empty) occured on aa20b238

2/ The server is rented

3/ I didn't knew there was an option to not use EOS ? We just used the default settings of the server from our hosting company. Where can I check this ?

4/ After removing the .ttp file, the player was able to join the server (but with a new, empty profile)

5/ In order to test a theory I had, I asked the player to move the game to another hard drive (uninstalling it fully, then reinstalling it on the main harddrive of the computer instead of an external hard drive). When he did that, he started to have some output_log again. Here they are:

Player output_log 1: http://gofile.me/5ijmE/eWDNuJgco

Player output_log 2: http://gofile.me/5ijmE/1jyKsjSyj

Please note that both of those output files were made AFTER that I deleted the incorrect file, so the player was able to login again - but I would like to avoid the file to become corrupted again as it's already the fifth time this player have to restart from 0.

In addition, the server's output_log is too big for me to put on pastebin (a bit more than 5Mb, where pastebin allow up to 512Kb). However, here is a link to download it in order to have a better view of it: http://gofile.me/5ijmE/rz5sLVfJA

 
Zæffy said:
Note for clarity: I realise that I may have done a slight mistake while translating in english ; since our last wipe, there is only one player who is still facing the empty .ttp file problem. In addition, every other players have output_log file (e.g. I do have them), it's only this specific player who doesn't.

1/ We always used the latest build, on the unstable branch when A20 was only on unstable, then on stable. The current problem (when the file went empty) occurred on aa20b238

2/ The server is rented

3/ I didn't knew there was an option to not use EOS ? We just used the default settings of the server from our hosting company. Where can I check this ?

4/ After removing the .ttp file, the player was able to join the server (but with a new, empty profile)

5/ In order to test a theory I had, I asked the player to move the game to another hard drive (uninstalling it fully, then reinstalling it on the main harddrive of the computer instead of an external hard drive). When he did that, he started to have some output_log again. Here they are:

Player output_log 1: http://gofile.me/5ijmE/eWDNuJgco

Player output_log 2: http://gofile.me/5ijmE/1jyKsjSyj

Please note that both of those output files were made AFTER that I deleted the incorrect file, so the player was able to login again - but I would like to avoid the file to become corrupted again as it's already the fifth time this player have to restart from 0.

In addition, the server's output_log is too big for me to put on pastebin (a bit more than 5Mb, where pastebin allow up to 512Kb). However, here is a link to download it in order to have a better view of it: http://gofile.me/5ijmE/rz5sLVfJA
Click to expand...


***************************************
1. reason i asked is i have seen and had people say they were using a20b238 experimental and when we went stable, they had issues... the a20b238 exp and a20b238 stable are identical with one exception - we just change the name to not be experimental. no game code was changed.

2. de-syncing between clients and dedis located (usually rented) machines is a pia and an ongoing effort to try to improve.

3. there IS NOT an option but there is a file in the base game directory where it can be disabled.. advice is not to because we need all tests to be with it on if we are to catch the issues and fix them.

4. its too bad to do that but at least he is able to play again... i dont know but seems the rented dedi servers should have backup option for just that type of issue.

5. someone sorta tested that already and claimed/indicated that HDD are more susceptible to corruption and de-sync because of read/write speeds compared to SSD and NVME M.2

Thanks  :)
QA Tester-unholyjoe
***************************************

 
Last edited by a moderator:

***************************************
@SylenThunder didnt you or meganoth have info on EOS and needing to install the current new distributable... i am seeing tons of EOS issue on server log and this may be up your ally.


Thanks  :)

QA Tester-unholyjoe

***************************************



 
You probably meant Sylen as I have only circumstantial knowledge from reading support cases.

A missing redistributable should normally lead to a message in the logfiles, but even a reinstall seems to have fixed errors on occasion!? Lets say it wouldn't hurt to reinstall "Microsoft Visual C Redistributables 2022" and check if that helps

And just to be sure, OP should ask the player with the problem if he did clean his settings like it is described in the blue banner link above

 
Last edited by a moderator:
The player has the connection issue to the EOS servers. Looks like the service is enabled and functioning just fine. No errors about needing C++ redis. 

I do find two parts of the log interesting though.

First we have 

Code: 
2022-01-03T22:28:48 1.564 INF [Platform] Initializing EOS
2022-01-03T22:28:49 2.205 INF [EOS] Initialize: Success
This is where we would usually see the error about it not being able to initialize because it's missing the redis.

However just a little bit later after loading the base data and server admin...

Code: 
2022-01-03T22:28:55 8.159 WRN [EOS] [LogEOSAnalytics - Warning] EOS SDK Analytics disabled for route [1].
What disabled it? What is route [1]?

Afterwards it seems to connect, but then looses connection.

Code: 
2022-01-03T22:28:56 9.249 INF [EOS] Login
2022-01-03T22:28:57 9.848 INF [GSM] Skipping me
2022-01-03T22:28:57 10.373 INF [EOS] Login succeeded, PUID: 0002cee7263b47c0a67432725ff78ca2
2022-01-03T22:29:47 60.096 INF [GSM] GameSparks Disconnected
2022-01-03T22:29:47 60.694 WRN [EOS] [LogEOSMessaging - Warning] Lost connection to Stomp on 'wss://api.epicgames.dev/notifications/v1/30b9e9e5f58b4f4e82930b3bef76d9e1/connect' () - will attempt to reconnect after 5.4 seconds. LocalUserId=[000...ca2]
2022-01-03T22:29:53 66.211 ERR [EOS] [LogEOSMessaging - Error] Failed to connect to Stomp! Could not initialize connection
2022-01-03T22:29:56 68.815 WRN [EOS] [LogHttp - Warning] 0000027BE0846A40: invalid HTTP response code received. URL: https://api.epicgames.dev/telemetry/data/datarouter/api/v1/public/data?SessionID=%7B43DB53E8-43CB-EC6A-B569-BA9E59E4CD89%7D&AppID=EOSSDK.PhaseRelease.ReleaseBuild&AppVersion=1.14.1-18059966%20-%20%2B%2BEOSSDK%2BRelease-1.14.1-CL-18059966&UserID=&AppEnvironment=Production&UploadType=sdkevents, HTTP code: 0, content length: 0, actual payload size: 0
2022-01-03T22:29:56 68.816 WRN [EOS] [LogHttp - Warning] 0000027BE0846A40: request failed, libcurl error: 6 (Couldn't resolve host name)
2022-01-03T22:29:56 68.816 WRN [EOS] [LogHttp - Warning] 0000027BE0846A40: libcurl info message cache 0 (Found bundle for host api.epicgames.dev: 0x27c15a436e0 [can pipeline])
2022-01-03T22:29:56 68.816 WRN [EOS] [LogHttp - Warning] 0000027BE0846A40: libcurl info message cache 1 (Hostname in DNS cache was stale, zapped)
2022-01-03T22:29:56 68.816 WRN [EOS] [LogHttp - Warning] 0000027BE0846A40: libcurl info message cache 2 (Could not resolve host: api.epicgames.dev)
2022-01-03T22:29:56 68.816 WRN [EOS] [LogHttp - Warning] 0000027BE0846A40: libcurl info message cache 3 (Closing connection 1)


I can only assume that this is a firewall or security software issue with the connection to the EOS server being blocked. I have not ret received enough information from those that keep having the issue after properly allowing it to determine a cause. My guess would be that it is still being blocked. In most cases though, allowing the firewall passthrough, excluding from antivirus and antimalware, and possibly whitelisting the EOS URL resolves the error.

I don't know what effect the client not being able to reach EOS has on playing with the server. 

 
Zæffy said:
I would like to avoid the file to become corrupted again as it's already the fifth time this player have to restart from 0.
Click to expand...
From their log file, it looks like Windows 10 needs updates and the GPU driver needs updates. Both are running versions over a year behind.

 
Last edited by a moderator:
@SylenThunder The warning about Analytics disabled for route (1) seems to occur in every log I’ve ever read, even in those with no EOS issues so I’ve just accepted that it’s a warning with no real concern similar to the fallbach handler couldn’t load library warnings. I’m sure it means something to the programmers though may not be important in figuring out issues although it would be interesting to find out what it means just for the knowledge.

 
Sorry that it took me so long to reply again, just wanted to make a last post to keep you updated and say that the problem is now fixed.

I followed @unholyjoe's suggestion and asked him to move his game install to another disk (an SSD one). Since then he had no more problem.

I still find it strange  that a problem on the client side (here, the game crashing because of HDD issue) can make the file on the server's side corrupted, but now that we know how to avoid that, it's good for us!

Thank you again to all of you.

 
You must log in or register to reply here.
Back
Top