Skip to content
Open
Changes from all commits
Commits
File filter

Filter by extension

Filter by extension

Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
11 changes: 9 additions & 2 deletions content/en/docs/tasks/administer-cluster/topology-manager.md
Original file line number Diff line number Diff line change
Expand Up @@ -271,7 +271,14 @@ lack of data, using this policy option with Kubernetes {{< skew currentVersion >
at your own risk.
{{< /note >}}

You can enable this option by adding `max-allowable-numa-nodes=true` to the Topology Manager policy options.
You can enable this option by adding to the Topology Manager policy options `max-allowable-numa-nodes=<numa_nodes_threshold>`,
where `numa_nodes_threshold` is an integer defining the new NUMA nodes threshold for starting the kubelet.
Here is an example of such configuration in kubelet configuration file:

```yaml
topologyManagerPolicyOptions:
max-allowable-numa-nodes: "32"

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

max-allowable-numa-nodes=<numa_nodes_threshold>, where numa_nodes_threshold is an integer — The old text said max-allowable-numa-nodes=true (boolean). This PR changes it to accept an integer. This is a significant behavioral claim — please link to the upstream KEP or source code commit that changed this from a boolean to an integer threshold.

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

If the value is an integer threshold, what does "32" mean exactly? Does it raise the default limit from 8 to 32? The text says "defining the new NUMA nodes threshold for starting the kubelet" but the existing paragraph at L276 says "Setting a value of max-allowable-numa-nodes does not affect the latency of pod admission". These two sentences are now contradictory in tone. Clarify: does setting "32" mean "allow kubelet to start with Topology Manager on nodes with up to 32 NUMA nodes"?
L267–L272 The note still says "nodes with more than 8 NUMA

```

Setting a value of `max-allowable-numa-nodes` does not (in and of itself) affect the
latency of pod admission, but binding a Pod to a (Kubernetes) node with many NUMA does have an impact.
Expand Down Expand Up @@ -394,4 +401,4 @@ assignments.
(beta) for more options.

1. The scheduler is not topology-aware, so it is possible to be scheduled on a node and then fail
on the node due to the Topology Manager.
on the node due to the Topology Manager.