I think these issues should be escalated. In the last 5 hours, over 50 servers have crashed according to my analytics I have been collecting.
That’s over 50 servers with up to 25 players playing.That’s 25 whole potentially lost customers for me, and not even half of those players are taking the time to rejoin/reconnect after being disconnected.
It’s hurting a lot of first impressions, and my game has been falling off charts because of constantly lost servers and players.
I think I have a fix for alexnewtron’s bug. It is a race condition in the BadgeService implementation - when you award a badge and the web call completes, it triggers a replication of an event that’s used to display the badge notification.
Unfortunately, it does this on a thread that is not synchronized with replication - as a result there is a race condition, that can cause an action-at-a-distance where we crash the server later when trying to send this event in a packet.
I am having crash problems with my game now, losing hundreds of players over it as it very frustrating for them to lose all their progress building up starbase resources. It tends to only crash after several hours (sometimes days) of play.
Could someone with log access give me some insight on what I did wrong?
imo this isn’t a good solution - having to manually get a staff member to tell you why your server crashed.
Is there no way for you guys to create a log without the sensitive information?
I completely scrapped a game and started over because I couldn’t for the life of me figure out what caused the crashed. I tried to fix it for nearly a month.
It’s really annoying having no way to figure out why your server crashes.
Sometimes, servers at Defend the Frontier experiences this, and I have no way to fix it.
I am also experiencing issues with users complaining that they are losing connection randomly in my game Territory Conquest.
I’ve experienced it myself and am at a loss as to how it is actually happening. I have tried monitoring lua memory usage on the server using collectgarbage(“count”) and found that it doesn’t accumulate after rounds. I make use of BadgeService, DataStoreService, and MarketplaceService but I’m not sure what would be considered a high enough frequency to cause problems if requests to those services are causing problems.
It would be great if the crash logs for my servers could be looked at…