Pretty often we get a lot of inquiries as to why Experimental/Unstable branch servers are not online, or requests to revert to the last playable build so people can play on the Experimental/Unstable branch servers while they’re offline. I thought I’d take a brief moment to try and explain exactly why when they are offline to the public they can’t just be turned back online.
As you all probably know, especially if you keep up to date with our Status Reports (Biweekly – found at devhub.dayz.com) the team has been working hard on resolving the current primary major blocker for testing 0.60 build on Experimental/Unstable branch – servers unable to accept client connections past 10(ish) players at a time. This is just one example of the many different type of type of issues that can and will come up during the development, testing, iteration, and regression of a new build in DayZ’s development cycle. “Just reverting” is far more of set back than just simply renaming a few files and pasting in the last known good build to the gameserver root folder.
– Each iteration on an Experimental/Unstable build entails more than just the executable and data folder for the game server
– Database structure, Servlet functionality (Central Hive), and Central Loot Economy are all also often iterated within this update process and are not often backwards compatible with older versions of the DayZServer software (In simple terms, the language changes a little bit each update and DayZServer.exe versions don’t often talk properly to newer versions of the database, servlet, and CLE)
– Internal testing for Experimental/Unstable updates occurs *on the Experimental/Unstable database/servlet* and if we revert, then we shut down the internal testing process on the next update
– Even if the public Experimental/Unstable branch servers are offline, we are often running internal tests on unlisted servers on the Experimental/Unstable environment in an effort to resolve any blocking issues and get the next update to Experimental/Unstable branch out there for public testing
I know I might have gone off on a bit of a tangent there, as I am often doing in these type of posts – but I thought some of you might like to know a little more why its not as easy as just rolling back a build when we’re working on getting Experimental > Stable.