87101b131a
When the scheduler assigns a device instance, it iterates over the feasible
devices and then picks the first instance with availability. If the jobspec uses
a constraint on device ID, this can lead to buggy/surprising behavior where the
node's device matches the constraint but then the individual device instance
does not.
Add a second filter based on the `${device.ids}` constraint after selecting a
node's device to ensure the device instance ID falls within the constraint as
well.
Fixes: #18112
|
||
|---|---|---|
| .. | ||
| benchmarks | ||
| annotate.go | ||
| annotate_test.go | ||
| context.go | ||
| context_test.go | ||
| device.go | ||
| device_test.go | ||
| feasible.go | ||
| feasible_test.go | ||
| generic_sched.go | ||
| generic_sched_test.go | ||
| preemption.go | ||
| preemption_test.go | ||
| propertyset.go | ||
| rank.go | ||
| rank_test.go | ||
| reconcile.go | ||
| reconcile_test.go | ||
| reconcile_util.go | ||
| reconcile_util_test.go | ||
| scheduler.go | ||
| scheduler_oss.go | ||
| scheduler_sysbatch_test.go | ||
| scheduler_system.go | ||
| scheduler_system_test.go | ||
| scheduler_test.go | ||
| select.go | ||
| select_test.go | ||
| spread.go | ||
| spread_test.go | ||
| stack.go | ||
| stack_oss.go | ||
| stack_test.go | ||
| system_util.go | ||
| system_util_test.go | ||
| testing.go | ||
| util.go | ||
| util_test.go | ||