Skip to content

ci: Don't run many more jobs than we have

Daniel Stone requested to merge daniels/virglrenderer:ci-concurrent-disaster into main

This explains a lot of stuff.

These jobs used to run on a dedicated 32c/64t machine. Now they run on a 32c/64t machine that's shared with up to 8 other jobs.

Stop trying to subvert FDO_CI_CONCURRENT so we don't massively oversubscribe the machine.

Sorry but I'm just insta-merging this; it's doing way too much damage. (Kicking myself I didn't spot it much earlier.)

Merge request reports