[Bug]: High node numbers result in delete messages and constant new nodes joining

[DooMMasteR]

Contact Details

No response

What happened?

I am and many others are currently at 38C3 and there is a very busy Meshtastic mesh on and around the event, my node alone saw ~48000 packages in a day.
With more than ~80-100 nodes, both the App and the node develop issues (the event is in excess of 200 nodes).
The first one is that they are constantly notifying of new nodes joining the mesh, the node (T-Echo) is limited to 80 nodes, but the app should be able to track more, it however seems to fail at it and constantly notifies me about new nodes I have 100% already seen in the past.
But even worse, the app lost multiple private chats with other nodes after forgetting the node.
Even my own relay (another T-Echo) and it's messages got lost this way.

App Version

2.5.15

Phone

Pixel 8 Pro

Device

T-Echo

Firmware

2.5.15

Relevant log output

No response

Code of Conduct

• I agree to follow this project's Code of Conduct

[infered5]

Can confirm this happens to me too. Using a RAK4631 and Galaxy Flip 3 (Android 14).
Meshtastic app 2.5.15 (Play store with Beta Tester access). Seems to happen around the high 80s of nodes in memory.

[jpryor223]

I am seeing the same issue as well, any nodes over 80 or 100 depending on which device are not loaded on reconnect. In particular I see nodes that are not encrypted (older firmware) mostly missing that were active in recent hours and all need to request info and position again. Other unset and unrecognized nodes are also not reloaded. The app node DB continues to have history for older nodes so once the node appears again I can see past logs, just takes time for info and position to update again like finding a brand new mode.

[jpryor223]

Also should note I am not seeing this issue from Apple app, that is maintaining more than >100 nodes on reconnect.