dbuild's cache is per-machine, so if the same community build runs on a different behemoth than it got last time, it doesn't get to use the cache from last time. the savings from the build cache is often vast, so this is definitely been worth doing. (I've been doing it manually for a while now by editing the config manually in Jenkins, though Chef comes along and resets it later.)
2.12 is where most of the action these days is, so I'm planning to pin both 2.11.x and 2.11.x-jdk8 to jenkins-worker-behemoth-1, with 2.12.x getting jenkins-worker-behemoth-2 all to itself. (well, except for the all the non-community-build jobs...)
dbuild's cache is per-machine, so if the same community build runs on a different behemoth than it got last time, it doesn't get to use the cache from last time. the savings from the build cache is often vast, so this is definitely been worth doing. (I've been doing it manually for a while now by editing the config manually in Jenkins, though Chef comes along and resets it later.)
2.12 is where most of the action these days is, so I'm planning to pin both 2.11.x and 2.11.x-jdk8 to jenkins-worker-behemoth-1, with 2.12.x getting jenkins-worker-behemoth-2 all to itself. (well, except for the all the non-community-build jobs...)