Jump to content

Liesel Weppen

Members
  • Posts

    1,433
  • Joined

  • Last visited

  • Days Won

    2

Record Comments posted by Liesel Weppen

  1.  

    Perhaps it's also possible to reproduce with just one player but two quests from different traders but in the same poi?

    If you have two quests in the same POI there is afaik no way to choose which one you actually want to start? Or does it matter which quest you have marked as active?

    In this case player 2 had his fetch quest active. The nighttime quest that wrongly started was not marked as active from player 1.

     

    I don't know how you implemented the quest handling. ;)

    But something seems to be fishy with having multiple quests in the same POI, at least if it is two different players.

  2. I didn't try to reproduce it, i just saw it happening. I don't even know how i can enforce it for testing and i for sure will not try and repeat accepting quests until two players have a quest in the same poi again.

     

    For reproduction i'd suggest to try:

     

    1) 2 Players on one server, both need to be "friends" for being able to share quests

    2) Player 1 accepts a quest of certain kind (fetch/clear/night), for example night

    3) Player 2 needs to get a quest in the same poi (however you can enforce this for testing), e.g. fetch

    3a) probably the quest of player 2 needs to be a different quest type, but not sure about that

    4)  Both go to the POI

    5) Player 2 (the player that accepted the quest in this poi as second) starts the quest at the exclamation mark

    ->

    From what i saw here it didn't start the fetch-quest of player 2 but instead the night quest of player 1 (during day).

     

    I didn't pay attention if in the quest log finally the mission of player 1 was completed or additionally wrong the quest of player 2.

     

    It's not on my server, but i asked the server owner for the logs. Probably something can be seen there, but it will not be "compact" as there were several hours of playtime between when each accepted his quest and when we actually startet that quest.

     

    As i said, we just noticed it because we had to repair a generator during day. The initial problem might even be older, but not sure if we would have noticed the different quest types if it just exchanged fetch and clear, because they are to similar.

    Probably additional hint, not sure if it maters: Both quests were from the same trader.

    Further information:

    Dedicated server was started with the first A20 experimental with of course a new map (random gen). The issue happend just yesterday with of course the current final A20.

    EAC is turned off, crossplay is turned off.

    Serverhardware runs in a virtual machine, serverhardware even changed in the meantime.

    But i'm quite sure, all of this is not relevant.


    If it might be relevant, the POI was a farm poi with corn in the front, a main building and a garage next to it, where the generator is in the garage. Iirc you follow the path through the main building and finally get into the garage over it's roof.

    I don't know the name of the POI.

    From the webmap it was this POI:

    grafik.png.23b7af61f34f09a59be8aa6a7a93a6e6.png

    Obviously a POI that can give night quests and fetch/clear. ;)

  3. Duh, since edit is not possible:

     

    Player A accepted the nighttime quest (long) before Player B accepted the fetch quest in the same POI. So i guess it's not per player, but queuing quests per POI in order you've accepted the quests at the trader?

     

    We often had quests at the same POI but since there have been no nighttime quests before A20, with the mentioned problems, we probably just never noticed it mixes up quests?

    Quest also started with the "power down" sound of a nighttime quest.

  4. 17 hours ago, marksisma_deliri said:

    But shouldn't programs or operating systems have integrity control mechanisms when reading/writing and handle such cases calmly?

    Nope, they don't have. They assume your hardware works correctly.

     

    17 hours ago, marksisma_deliri said:

    Or such mechanisms work only with file system, but not with RAM?

    Neither harddisks nor RAM have such mechanisms by default.

    For RAM you need at least memory modules and mainboard/cpu that support ECC. And that can only correct certain types of errors but not everything.

    For harddisks you need to put up a raid or similar yourself.

     

    I also don't know the exact meanings of the errors, but if memtest prints such errors, your RAM might be broken.

  5. malloc(): invalid size (unsorted)

    from your log points to some problem with memory allocation. Since your video shows you have 16G RAM it should be sufficient.

    If your installation is corrupted, try verifying your files, or even better download it from scratch. If possible also try installing it to another hdd/ssd just to test if eventually your hdd/sdd is broken.

  6. @Elbobburrito

    since it is a timeout problem while downloading the world and you said they are already reaching 80%, increasing the transfer speed might be enough.
    By default the download speed is limited to iirc 512kb/s (4Mbit/s) in serverconfig.xml. Of course your servers upload bandwidth and your friends download bandwidth needs to be faster than this to make it work.

     

    I don't know if 7d2d is capable of resuming the world download or if it starts from scratch if you try again. In theory your friend already has received 80%. If the download is able to continue, simply unban them and let them connect again to download the remaining 20%.

    @Jugginator

    How long is the timeout until ddos protection kicks in?
    For an 8K world obviously it was longer than it took to download the world with the above mentioned default speed limit. A 14k map is 3 times the size, don't know if that applies 1:1 to the file size. However 1k maps are officially unsupported anyway?

    @Elbobburrito or if it is a private server and this problem is caused by EAC, simply run the server without EAC. Or run the server without EAC, let your friends connect and download the world and then enable EAC on the server again, if required.

    For the EOS thing it probably helps if you set "crossplatform=None" in platform.cfg, if all of you using steam anyways.

  7. @TSBX: Whats "blow" shows: Sorry, we can't show this content because you do not have permission to see it.

     

    High RAM usage (under Linux) was already reported. From what RAM usage i saw when playing the game, 7d2d alone used up to 14GB during play (world generation even peaked at 22GB). Depending on what other stuff is running in background, 16GB might not be sufficient to run the game (at least without needing swap).

  8. Had some further tries. Was not able to join the server when using vulcan. It did not crash, but completely hangs while joining (i killed it after several minutes of waiting).

    In singleplayer vulcan works.

    No idea how using vulcan for graphics could affect joining a server, but obviously it does.

     

    Are there any details on what "not fully supported" means? What could we expect, what not?

  9. On 12/8/2021 at 4:20 AM, chibbs said:

    but joining a server it crashed several times for me.

    Just an indication, but but joining a server while using vulcan also crashed my client. Since it was the first join to this server it also broke my starter quest. In the second try i joined with GLCore, but had no starter quest, but at least it worked. Since it was my first join anyway, i deleted my char on the server and joined again (with GLCore) and it worked from this point.

  10. Same here, didn't see the circle at all with GLcore. Haven't tried vulcan yet, because the launcher says "not fully supported".

     

    At least you can see if you are inside the unvisible circle by the flashing red chest symbol on the compas.

     

    Also noticed very high memory usage under linux. Was at ~18GB JUST for 7d2d, when i exited the game.

  11. Can confirm. And most likely i figured out:

     

    Our dedicated server is running on a Phenom II and is also crashing with Sigill on libEOSSDL-Linux-Support.so.

     

    Server runs fine on Linux with a ryzen 2700x and ryzen 5950x.

     

    Finally after hours of googling, a similar bugreport for Satisfactory dedicated server also trying to run on a Phenom OO got me in the right direction.

    I disassembled the libEOSSDL-Linux-Support.so and it turns out it contains AVX instructions. And the Phenom II does not support AVX.

×
×
  • Create New...