Wait so are you saying that there will be 3 SEPERATE servers that hold 3 smps? Or are you saying u will be adding 3 smps?
They went and took this picture and sent it to me.... Yes still with same host, I really like them To the users like yourself, nothing is changing. Right now we have 5 servers purely to host the game servers, and were lowering it down to 3 (but not going to cause lag because they are better servers) Maint is in roughly 1 hour. I'm running a script every 5 mins that syncs the data from old to new server, so itll be very little to copy once the servers shutdown, to keep the downtime as minimal as possible. https://gist.github.com/aikar/6cc08e9a880fcac5f569
There is too much that needs to occur within the minecraft multiplayer development community before we can update to 1.8 for us to give even a vague answer.
Getting close. I was reminded of another thing I've been wanting to do, to setup a big system configuration called Huge Pages. This WILL provide a massive performance boost, so I am finalizing setting that up to make sure the servers are running with the best possible speed.
For some reason i have never experienced lag Even at the He's a Pirate song in the cow hockey arena. Maybe I'm just getting all the processing power.
the default server is constantly rotating every 5 minutes usually. You logged out on smp2 5 hours ago is why you logged back in there.
Also yes, servers are now back up! Everything went as planned. No issues occured (other than me having to play around with Huge Page settings to get it working) EMC is now even more redundant, with Bungee running over 4 servers now randomly being used, so bungee going down can no longer take us down completely ==== Summary of changes! ===== 1) Moved from 5 Dedicated Servers powering the Game Servers to 3. - Newer servers are much more powerful and able to handle more load than the previous. - Now running 3 SMP's per server, with Utopia/Stage/Games residing on MasterDB - Still have 3 other servers: DB, WebProxy and Web 2) Moved to a newer SSD drive. - Reset the timer for failure. SSD's fail a lot faster than HDD's, and MC and the like would reduce the life span even more from constant writes. Since those old drives were nearing 1.5 years old, they were getting to a risky age. - Newer drive with faster specs than previous 3) Enabled "Huge Pages / Large Pages" - Lets the OS / Java access memory MUCH more effeciently, giving a substantial performance boost 4) Spun up 3 more BungeeCord instances over the 3 new servers, and Round Robin them in. - Splits up load across multiple servers - Redundancy incase 1 crashes (and for some reason doesn't auto restart. - Resistance to network level issues on a single IP (multiple IP's) 5) Database Server Upgrade - Updated to Percona 5.6, the latest stable with many changes to the system - Enabled some needed config changes to alleviate DB issues that we've seen some times. - Adjusted memory settings - DB was using too much which left less room for the other stuff, which made us have to leave stage offline for a while. (I was EXTREMELY generous before on memory not expecting to run other apps on the server, way more than was needed to load our entire DB into memory... which is unnecessary) Quite a lot of work... Time for me to sleep now.
There was a downtime in voting this morning, but I resolved it before I went to work. Forgot to open the votifier port on firewall.
