-
Notifications
You must be signed in to change notification settings - Fork 31
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
ci: fix dunfell github actions #49
Conversation
Containers build quicker than a full QEMU as no kernel etc. is needed
@bachp friendly ping on this one |
@agherzan Unfortunatly github actions doesn't work well with yocto builds. I haven't found a reliable way that doesn't run into timeouts all the time. Ideas are welcome |
@bachp I have finally migrated to GitHub actions with a BSP layer. I will test drive the setup for a while but it looks like everything is in place for now. You can take a look. I'm planning to define some yocto-specific actions so we can reuse bits across the ecosystem. |
@agherzan I have CI working via kas, but it still takes a lot of time. Ideally there would be a way to cache some of the sstate between runs. |
@bachp not easy if you use default workers. I'm not sure what setup you have here. |
@bachp I noticed you got a successful build with CDN SSTATE_MIRROR for mickeldore. I was contemplating if the 10GB limit on cache@vx would also be enough to help us keep under the dreaded timeout. |
@moto-timo Unfortunately I'm not happy with the any of the current CI setups I tried. I haven't tried cache@vx but so far I found that the sstate doesn't fit well with the method of uploading and downloading the cache. As the cache keeps growing. But you got me thinking about an idea. Maybe a workflow that would work is the following:
I haven't tried this yet. |
Obsolete now |
This fixes github actions for the dunfell branch