Hi folks, We are planning to release a new feature (called "Slave Restart") that allows organizations to upgrade mesos without disrupting the running tasks. Previously, when the slave process was stopped for an upgrade, the tasks running on that slave committed suicide. With the new feature, the tasks can keep running and connect back to the upgraded slave, when it restarts.
Since this feature includes a huge amount of code change, we would like to create a new branch (0.13.x). The new branch helps us to commit bug fixes easily, test the feature and also unblocks people from committing other major changes/features to master/trunk. We have been testing "Slave Restart" feature for a while, but need some more time to iron out some kinks. Note that, any bug fixes that relate to "Slave Restart" will go to both "master" and "0.13.x". Let us know if you have any questions/concerns.
