Loading into team create within our developmental test place for, Grand Piece Online, ends up erroring and not letting you into team create 99% of the time, this happens to multiple people who’ve tried. Even after successfully loading in, sometimes team create boots you out of the team create and you have to re-try all over again (I’m assuming the team create server crashed?). Video of the problem attached below and logs attached in report.
Images:
Incident IDs in order: 1895737653260431375, 201787829750348748, 883725142411384570
System Information: Windows 10 Home, i9-10850K, 64 GBs of RAM, RTX 3090
Repro files not provided as it’s the team create of the place that fails to load. Without team create enabled you can load in fine.
Expected behavior
Team create should properly load every time rather than having to attempt to load team create 100 times. Team create shouldn’t also crash after loading in and working for a few minutes.
The place loads without Team Create, however, without team create we cannot have multiple developers work on the game without it and that’ll cause some issues with our development timeline.
Publishing the place to another empty place will only load in if it doesn’t have team create enabled, having team create enabled will cause it to no longer load in. TL;DR the issue is only with team create on. (Link to private game that we published over: test 2 - Roblox )
I’m not sure on what you mean by “offload some instances or some of your memory”. We could use InsertService to load our maps into the game, but other than that there’s not much we can offload from out of studio realistically and that has some caveats like if Roblox’s services go down the map would be unable to load for some time while attempting to pcall and retry with a delay. For now, we’ll try to delete some unneeded stuff like old assets and testing assets for now to try to mitigate the issue.
I also develop for another front page game called Peroxide that has a similar file size to Grand Piece Online but the team create hasn’t had any issues unlike Grand Piece Online. (Peroxide is 33 MBs and GPO’s is 42 MBs)
Thanks for the follow-up. What’s happening we believe is that you’ve reached our Team Create RCC memory limit. Could you do a quick check using the Performance widget in non-TC Studio to see how much memory is being used by your place? If it’s ~10gb you’ll start to get this error
I looked into the performance widget in team create and it appears that the thing using the most memory is CoreMemory → studio/undo, which is using more memory than everything else combined in the studio instance.
We tested on multiple computers, accounts and with all plugins disabled and seems to be caused by the engine itself. Any clue on how to either mitigate this or what’s causing it? Without using control z/undo/redo I don’t think it’s intended for studio/undo to be using nearly 7 GBs of memory.
I wasn’t able to take the picture while team create was on, if I turn team create on I cannot load into the team create as it fails, and even if I load in it disconnects you with a failure message in the original post.
Gotchya - that undo/studio memory issue is a problem, but actually doesn’t exist in RCC TC. In other words, that won’t cause your TC session to fail. Can you send a snapshot of the memory profile using F9 from an empty game server? Maybe from the hidden published place?
How do I check how much memory the RCC TC is in usage and what’s using it, I’m able to load into it for about 30 seconds before it disconnects us sometimes (Takes a lot of tries though), when I checked the performance widget while in team create it showed the same CoreMemory undo tab using all that memory again.
Thanks for screenshots! Sorry for late replying was digging into your error. Your f9 shows in-game memory usage - so that’s normal. The Performance widget (which needs to be updated) only shows Studio client memory usage (so not TC), and so the Undo Stack will not be the reason TC session is crashing most likely still - because that’s showing your client Memory.
I need some more crash logs. Could you try opening your place in TC again and sending me the new incident IDs (and place ID just to be safe)? or if you’re comfortable you can send me the place file via DM or the debug group, and I can investigate through there too.
I can’t access a specific studio experience at all. after dozens of attempts and restarting my PC, I get the same error message.
The place is nothing but a storage for assets because otherwise roblox studio can’t handle all of them in one place without failing to render them. This one is particularly small though with only a couple hundred on a baseplate so it shouldn’t have much memory usage to begin with.
Hey all,
we had a nasty bug with audio here that was causing memory issues, should be resolved now. If anyone still has issues with audio and studio crashing let us know. Else we’ll close this thread out.