999 ping only on one player since 13.0 update#474
A friend we play with regularly who before had ~100-150 ping is now almost the entire time playing with 999 ping to the host. He is the only player in our group to whom this happens.
We’ve found a strange workaround for now - I turn on my VPN and it’s gone. I feel I should mention our countries, it might matter.
Original connection is: USA(player) -> Germany(host). Through VPN I switched to a server and hosted in the Netherlands.
The issue may be related to the network routes between cities and countries. Report to your provider and to the friend’s provider.
Is this happening on all tables, or a specific one?
If you haven’t checked, could do run this bizarre control experiment? What if you load a clean “nothing-on-it” table?
It sounds bizarre.. but I’ve seen this problem occuring on one table, and then disappearing on a different one, so perhaps it’s somehow linked to how/where-online various assetts are being pulled from?
It’s related to this, I’d assume:
https://tabletopsimulator.nolt.io/471
If it’s the same, it’s not a network routing issue, as keeping the same host but changing to another table fixes the issue.
^ that’s what I was curious about, does that IF apply?
We did not think to try a different table (we all just play one game hah). We have not had the problem for 2 days now, It was happening all the time before that. If it happens again we’ll try a different table.
I see this has been archived.. so just in case I wish to report that the ping issue appears to still be there in certain conditions. The whole thing is quite bizarre and tough to pin down, but it’s real.
https://tabletopsimulator.nolt.io/471, linked above appears to still be an active issue.
After some discussions among the table/script makers of the tables linked in #471, we came up with one hypothesis:
- Perhaps it’s based of collision calculations? The ping issues arise with a custom table, hosting the custom assetbundles elsewhere does not make a difference. However, switching to a level with TTS’s vanilla table model seems to remove the ping issue entirely. Removing many/most of the background elements on the high-ping-custom-model level seems to delay the ping spike for some time. So, the hypothesis is that the custom assetbundles have collision boxes that end up causing an infinite/long loop of collision calculations on the host’s computer, and the sending of these calculations to someone on a different continent results in the ping spike? .. or who knows why. By this point, the involved players have simply given up on the high-ping table and don’t use it anymore.
Reinforcing the previous point by Dmitry-
I’ve helped with the testing mentioned for #471; my theory for the consistent 999+ ping is that something like the aforementioned collision bug causes a large lag spike, that can last multiple seconds. The ‘infinite’ ping issue happens if the same event triggers another time before the first spike passes, and then again and again… causing a ‘ping leak’ of sorts.
Reducing the amount of elements on the table lessens the data being transferred, making the initial spikes last for a shorter period, therefore lowering the odds of them overlapping with a second event.
In Discord, the OP has confirmed this is not just a display glitch or a problem with the ping code; the ping accurately reflects the performance received,