Redhat has a great blog post describing CPUShares, but it assumes processes reside in leaf CGroups, and doesn't discuss how CPU time is calculated for processes that reside in branch CGroups.
For example: what is the CPU usage for each of these processes?
- CGroup A (CPU Shares: 1024)
- Process 1 (spinning at max CPU)
- CGroup B (CPU Shares: 512)
- Process 2 (spinning at max CPU)
- CGroup C (CPU Shares: 256)
- Process 3 (spinning at max CPU)
- CGroup D (CPU Shares: 128)
- Process 4 (spinning at max CPU)
| Process | CPU Usage (%) |
|---|---|
| Process 1 | ??? |
| Process 2 | ??? |
| Process 3 | ??? |
| Process 4 | ??? |
I'd love it if someone could help clarify this for me!
