It would be nice if bitbake supported splitting the builds onto multiple different machines to allow the load to be distributed. With the advent of multiconfig, this becomes something which is more desirable. It likely won't affect normal builds times too much due to the cross compiler creation still being the bottle neck in the build process, parallelisation doesn't help this. The logical place in bitbake to handle this is at the bitbake-worker level. This communicates with the server over a pipe so doing this from a remote machine shouldn't be too different. With multiconfig we already support spawning multiple workers. This bug has recipe specific sysroots as a prerequisite.
We now have recipe specific sysroots and working multiconfig. To make this work, we'd probably assume an NFS shared sstate that all the workers can pool sstate to and force all tasks for a given WORKDIR to run on a given host. The challenge would be ensuring multiple populate sysroot tasks execute on each worker to ensure the sstate is extracted.
Bulk move from 4.99 or 0.00 to 5.99
Do we just admit we're going to close this?