Closed hnez closed 5 months ago
Hi @ejoerns,
it looks like the GitHub action ran out of disk space (at least according to bitbakes disk monitor).
Can this be fixed by providing appropriate sstate via the sstate mirror so that less compilation takes place in the action?
Or should we change the BB_DISKMON_DIRS
to reduce the limit to below 1G?
Or move the rust-llvm-native
build into its own bitbake call before the autobahn
build so that its rm_work is executed before the python builds start?
Hi @ejoerns,
it looks like the GitHub action ran out of disk space (at least according to bitbakes disk monitor). Can this be fixed by providing appropriate sstate via the sstate mirror so that less compilation takes place in the action?
Yes. I have re-triggered my failed sstate server job. Let's see how far it comes..
Or should we change the
BB_DISKMON_DIRS
to reduce the limit to below 1G?
I wouldn't assume that this makes a notable difference
Or move the
rust-llvm-native
build into its own bitbake call before theautobahn
build so that its rm_work is executed before the python builds start?
Could work, but let's try updating/fixing the mirror first.
I see scarthgap sstate doesn't work anyway since the sstate job fails due to layer incompatibility. I'd tend to say we just merge this and check for possible fallout later.
I would have preferred to see the job succeed before merging, but would be okay with merging it as-is. I do not have permissions to merge this so if anybody else wants to do the honors feel free to do so.
I guess the correct youth term here is "YOLO".
Recently CI jobs have started failing since poky master has changed its LAYERSERIES_COMPAT to scarthgap.
Add scarthgap compatibility and while at it remove mickledore, which has reach its EOL.