Closed lefnire closed 11 years ago
Alas my friends, between this and the memory leak I don't think I'll be using Derby much longer. I've been grinding my gears too long on issues and not programming features.
Hey,
Hang in there! Or at least keep an eye on this. Seems like there has been activity recently on racer & derby, so who knows what will happen soon. Maybe only progress? We'll see. We'll hope.
But I don't know, I agree that if you want to finish a project now, maybe looking at some other possibilities isn't a bad idea...
No longer relevant in 0.5
Running habitrpg - we have many concurrent users at any given time, and due to a memory leak (still don't know if it's Racer, derby-auth, or habitrpg) we have to kill and restart the node service periodically. Every time that happens, Mongo spikes about 1k-1.5k queries per second for about 2-3 minutes, and I have to restart the service a second time to "calm it down".
In my limited understanding, I'm assuming all clients are trying to reconnect their Socket.io session to Racer at once causing an overload, and that maybe the second restart officially kills the connection - telling everybody to start over?