Open manzarek123 opened 2 years ago
At this size/bps it's likely triggering a lot of immediate resends (even worse if any transmission actually dropped fragments).
Sadly, this isn't a very reproducible scenario, so this would need someone who does have the ability to work on and test network stuff like this to fix it.
I can confirm this was a problem with just a 100 players. TriggerLatentClientEvent was replaced with just TriggerClientEvent and the issue has stopped.
Instead of 'confirming' and not providing any additional info to make this actionable, how about providing the info asked for, or testing with less nonsense bps values, or whatever? 😕
Instead of giving a snarky sarcastic reply that does nothing but encourage the exact opposite how about you actually ask for more information in your previous post instead. I don't see any. If I had more information to give then I would... I was simply confirming this isn't an issue with just one person so maybe it could be investigated more. BTW it was tested with just 250, it made no difference.
Sigh. Another issue lost to the trolls that start demanding and arguing in weird non-constructive ways.
If this issue were able to be investigated, it already would be, and probably would already be fixed. 'Confirming' to show 'please investigate!' doesn't add anything of note, and it only makes for more annoyance since there isn't any means for us to investigate this.
We don't have access to any server with even 5 active players, let alone your 'just' 100 players, nor do we have access to your code, or any of the other setup involved in replicating this. That is also what the initial response was pointing at: this requires such specific context to be able to 'investigate' and fix, that such a fix can only be done by an affected server owner who can test stuff themselves.
Again, 'same issue' style responses do not help. An issue does not magically become more actionable because more people are reporting it. We don't discriminate issues based on 'how many reports there are', and rather, the more reports there are for a non-actionable issue without reproduction steps, the more likely it is to be ignored as it becomes a sensitive subject to even think of touching.
When using TriggerLatentClientEvent to a lot of players (tested with ~1000 players) the network thread will freeze and everyone will timeout and be disconnected.
The same event to everyone with TriggerClientEvent works fine. The same event with TriggerLatentClientEvent to a single player works fine as well.
Event data was 47047 Bytes and bps was set to 1000000: TriggerLatentClientEvent('eventname', -1, 1000000, data)
I have and ETL trace, not sure if useful for this issue Trace might have unusual activity as a lot a players where reconnecting when I provoked the crash for the ETL record.
https://1drv.ms/u/s!Asx7IOhJxcoWhZZSCwUrWR_YJTBcAw?e=OcjxPz