05-10-2015, 05:04 AM
I've made changes just today in hopes to assist in resolving these intermittent connection issues further.
Unfortunately instances cannot actually communicate with each other, and a medium storage has to be used for communication, and it can be difficult to do perfected integrity checks to ensure two instances are not running at any point in time. There's been problems with ghost instances; server instances that are still running, but I can't even see them, close them, stop them, etc that continue processing a player as they are left behind. This is pretty rare, but it can be difficult knowing whether that is the ongoing issue for some players, or if it's just an issue in the integrity checks.
Part of the problem is out of my control (between goofy instances, end-user intermittent connectivity, dns routing issues, etc), the other is a problem with the work around methods. This is an ongoing thing, and I intend to keep working on it until it's as stable as possible, but the reports recently are very helpful as to what areas I can start covering and improving.
I hate seeing so many people with the connection issues, hence why I wrote the topic about not just telling me, "I got disconnected twice!", because all that does is frustrate both of us, but with the details as of late, I can actually work towards finding the issue that I personally haven't been able to reproduce through test accounts or on the development cluster.
Ray
Unfortunately instances cannot actually communicate with each other, and a medium storage has to be used for communication, and it can be difficult to do perfected integrity checks to ensure two instances are not running at any point in time. There's been problems with ghost instances; server instances that are still running, but I can't even see them, close them, stop them, etc that continue processing a player as they are left behind. This is pretty rare, but it can be difficult knowing whether that is the ongoing issue for some players, or if it's just an issue in the integrity checks.
Part of the problem is out of my control (between goofy instances, end-user intermittent connectivity, dns routing issues, etc), the other is a problem with the work around methods. This is an ongoing thing, and I intend to keep working on it until it's as stable as possible, but the reports recently are very helpful as to what areas I can start covering and improving.
I hate seeing so many people with the connection issues, hence why I wrote the topic about not just telling me, "I got disconnected twice!", because all that does is frustrate both of us, but with the details as of late, I can actually work towards finding the issue that I personally haven't been able to reproduce through test accounts or on the development cluster.
Ray