On Thu, May 5, 2016 at 12:20 PM, Marcelo E. Magallon <marcelo.magal...@hpe.com> wrote: > On Thu, May 05, 2016 at 06:45:36AM -0700, Khem Raj wrote: > >> > This works: >> > >> > $ MACHINE=qemux86 bitbake core-image-minimal >> > $ MACHINE=qemux86-64 bitbake core-image-minimal >> > >> > and bitbake parallelizes the build across tasks. What I'm trying to >> > figure out is if it's possible to run two bitbake instances with >> > different MACHINE values on different hosts but writing to the same >> > workspace (say, over NFS). >> >> Thats not possible and its by design. > > Can you nudge me in the direction of some documentation for this design?
It was not a conscious choice to limit however that is what it started with single machine singe distro in one invocation. It has been raised in past and I was one of the persons who raised it as a limitation if you want to call it. However in your case you said NFS and we have seen problems with NFS even with single machine builds in past so just be aware. Secondly during build there are native packages built and staged and every machine will build them similarly there are all-*-* packages which are going to conflict if two bitbake instances did not synchronise, they might conflict with each other, so unless we change the logic to let bitbake take machine as parameter and setup global datastore accordingly e.g. bitbake machine1,<image1, images2,.,,, machine2,<imageA, imageB....> it might not work. > > Thanks! > > Marcelo -- _______________________________________________ yocto mailing list yocto@yoctoproject.org https://lists.yoctoproject.org/listinfo/yocto
