Like the current device plugin interface, a DRA driver using this model
announces a list of resource instances. In contrast to device plugins, this
list is made available to the scheduler together with attributes that can be
used to select suitable instances when they are not all alike.
Because this is the first structured parameter model, some checks that
previously were not possible, in particular "is one structured parameter field
set", now gets enabled. Adding another structured parameter model will be
similar.
The applyconfigs code generator assumes that all types in an API are defined in
a single package. If it wasn't for that, it would be possible to place the
"named resources" types in separate packages, which makes their names in the Go
code more natural and provides an indication of their stability level because
the package name could include a version.
When a claim uses structured parameters, as indicated by the resource class
flag, the scheduler is responsible for allocating it. To do this it needs to
gather information about available node resources by watching
NodeResourceSlices and then match the in-tree claim parameters against those
resources.
The assume cache in the volumbinding plugin can be created with no separate
index, but List then failed because it tried to use the empty index name
instead of using the store's List function.
Blocking API calls during a scheduling cycle like the DRA plugin is doing slow
down overall scheduling, i.e. also affecting pods which don't use DRA.
It is easy to move the blocking calls into a goroutine while the scheduling
cycle ends with "pod unschedulable". The hard part is handling an error when
those API calls then fail in the background. There is a solution for that
(see https://github.com/kubernetes/kubernetes/pull/120963), but it's complex.
Instead, publishing the modified PodSchedulingContext can also be done
later. In the more common case of a pod which is ready for binding except for
its claims, that'll be in PreBind, which runs in a separate goroutine already.
In the less common case that a pod cannot be scheduled, that'll be in
Unreserve which is still blocking.
This moves adding a pod to ReservedFor out of the main scheduling cycle into
PreBind. There it is done concurrently in different goroutines. For claims
which were specifically allocated for a pod (the most common case), that
usually makes no difference because the claim is already reserved.
It starts to matter when that pod then cannot be scheduled for other reasons,
because then the claim gets unreserved to allow deallocating it. It also
matters for claims that are created separately and then get used multiple times
by different pods.
Because multiple pods might get added to the same claim rapidly independently
from each other, it makes sense to do all claim status updates via patching:
then it is no longer necessary to have an up-to-date copy of the claim because
the patch operation will succeed if (and only if) the patched claim is valid.
Server-side-apply cannot be used for this because a client always has to send
the full list of all entries that it wants to be set, i.e. it cannot add one
entry unless it knows the full list.
When dealing with unschedulable pods, the intent was to deallocate only claims
which are allocated and use delayed allocation. That if check wasn't handled
correctly, causing also claims with immediate allocation to be considered as
candidates.
Found during code reading, probably has never occurred in practice yet.
