by 0xd4f1cab694c4424c4796549edbb9b489789f4df5 (TudaMoon)
Servers can’t handle more than a given amount of users on a scene without crashing. This is causing major issues with the overall growth of Decentraland. If this issue cannot be fixed, then I am waiting for other options like the Otherside to come out who has proven in Beta to be able to handle thousands of people at once. We cant even handle double digits. Please vote yes if this needs to be implemented.
Between this issue, and the previous bugs when changing wearables and it not pushing the update, people are losing interest in DCL. These issues need to be fixed and soon, the upcoming Music Festival will be a disaster if not.
I agree, IT infrastructure & servers need to get repaired/improved ASAP. Even at low and scene radius 1 at a party of 40 is causing me to crash (which discourages me to make content with low graphics). I’m not sure how we got through Fashion Week, but the performance of DCL’s servers the last few weeks have been worse than Fashion Week, and were talking only 500-700 people active at a given second compared to 4000 during FW.
Lets prioritize further IT development ASAP.
I down but I don’t see the solution here just see the issue. Let me know when there a solution to vote on and I down to vote for it. thanks bro.
We need to allocate emergency funds to fix the backend infrastructure ASAP!
To whom? You know it will have to go through the proposal process. So until that time, other than this drawing attention to the matter it fixes nothing.
“Throwing money at a problem does not fix it” - ME
If this even needs to be a question, Decentraland is in trouble.
Lots has changed since DCL launch in tech, this should be priority #1 over all and everything.
All that money in the treasury and there’s a server issue?
Why
Well they don’t have a way to propose infrastructure changes. They are centralized in that way.
I heard the otherside is filled with a bunch of bored apes tho. that’s no fun.
This is the only process they have. This is there stupid process to get infrastructure changes. It has to pass this first apparently, then it can be actually voted on.
I can vote yes on this proposal with my miniscule VP. I can attempt to influence those with more VP to vote yes as well. However, until there is someone that steps forward and says they can fix the problem and proposal is drawn up and passed then this is a waste of time. Sorry but it truly is. I totally get the frustration cause I avoid parties often because of it. The fact remains that “You can throw money at a problem but that does not fix it”
This does bring to light there needs to be an “Emergency” or “Urgent” or “Critical” type proposal option to take care of something that is one of those categories.
Do you not believe in Decentraland? Then why are you here?
Regardless of whether or not throwing money at a problem fixes it does not matter. What matters is that this is Urgent and Critical to DCL’s success and if the DAO does not start throwing money at it the result will be bad.
The masses are coming. If it’s not resolved pronto the term will be “Don’t have parties or events in DCL unless it’s small.”
Yeah I agree, no expression there, but we should easily be able to handle larger masses of people.
Hey folks, I’d be happy to help with this issue but it is the first time I hear of it.
Can you help me with a more detailed description of what is happening? Servers seem super stable… It may be something somewhere else… I’ll need descriptions, videos, reproduction steps, anything to know where to look …
Thanks
While I agree this is a problem that needs to be solved and is extremely common for any large events, I don’t believe this is a server issue. This may just be an issue with browser limitations on the amount of accessible memory.
@menduz The error in large scenes is typically the following:
Cannot enlarge memory arrays. Either (1) compile with -s TOTAL_MEMORY=X with X higher than the current value 2139488256, (2) compile with -s ALLOW_MEMORY_GROWTH=1 which allows increasing the size at runtime, or (3) if you want malloc to return NULL (0) instead of this abort, compile with -s ABORTING_MALLOC=0
For large events with streaming video and moving scene objects we may need to separate it from other large areas on a scene and potentially add and remove scene objects when entering the venue area.
The streaming quality settings also seem to have an impact on crashing. When streaming in 1080p with a high bitrate and it was impossible to use the scene. When using 720p with 2500kbps it was much more stable.
Ideally we solve the issue and make things more stable, but in the meantime may need to be more conscious of how much resources a scene is using and manage that properly in larger scenes.
Do you know which scene it was on? Can you provide the coordinates please?
@HPrivakos it is happening at -147,133 Decentraland
Thanks.
Was the issue there since several days or only since today/yesterday?
Is it happening on other scenes too?
