Update cluster-resources.md #736
Conversation
karmada-bot
added
the
kind/documentation
karmada-bot
added
the
size/S
| @@ -267,7 +267,8 @@ The above is a cluster resource model with three grades, each grade defines the | |||
|
|
|||
| #### Schedule based on Customized Cluster Resource Models | |||
|
|
|||
| `Cluster resource model` divides the nodes into levels of different intervals. When a Pod needs to be scheduled to a specific cluster, `karmada-scheduler` compares the number of nodes in different clusters that satisfy the requirement based on the resource request of the Pod instance, and it schedules it to a cluster that satisfies the requirement with a larger number of nodes. | |||
| `Cluster resource model` classifies the nodes into different grades, which are essentially resource ranges. When a Pod needs to be scheduled to a specific cluster, `karmada-scheduler` compares the number of replicas able to fit on each cluster. To do that it adds up | |||
There was a problem hiding this comment.
I like the word classifies. Thanks.
| @@ -267,7 +267,8 @@ The above is a cluster resource model with three grades, each grade defines the | |||
|
|
|||
| #### Schedule based on Customized Cluster Resource Models | |||
|
|
|||
| `Cluster resource model` divides the nodes into levels of different intervals. When a Pod needs to be scheduled to a specific cluster, `karmada-scheduler` compares the number of nodes in different clusters that satisfy the requirement based on the resource request of the Pod instance, and it schedules it to a cluster that satisfies the requirement with a larger number of nodes. | |||
| `Cluster resource model` classifies the nodes into different grades, which are essentially resource ranges. When a Pod needs to be scheduled to a specific cluster, `karmada-scheduler` compares the number of replicas able to fit on each cluster. To do that it adds up | |||
| the products of the `#(nodes)` * `#(replicas that can fit in each node)` and schedules the Pod to the cluster that can fit the most of them (highest cluster sumproduct). | |||
There was a problem hiding this comment.
That may not be correct, the cluster resource model is used by schedule to figure out how many replicas can hold for a specific cluster. That's the only mission.
There was a problem hiding this comment.
So, would you suggest to delete the part starting from "To do that ... "? I based the conclusion on the examples below, where the cluster that could most pods was chosen.
There was a problem hiding this comment.
Some like:
The Cluster Resource Model classifies nodes within a cluster based on their available resources. This classification helps the scheduler determine the number of replicas that can fit into each cluster. By categorizing nodes according to their available resources, the Cluster Resource Model enables the scheduler to make decisions about where to place application replicas based on different scheduling strategies.
| @@ -351,24 +352,24 @@ status: | |||
| grade: 6 | |||
| ``` | |||
|
|
|||
| Suppose the Pod's resource request is for a 3-cores CPU and 20Gi of memory. All nodes that meet Grade 2 and above fulfill this request. Considering the number of nodes available in each cluster, the scheduler prefers to schedule the Pod to member3. | |||
| Suppose the Pod's resource request is for a 3-core CPU and 20Gi of memory. All nodes that are classified as grade 3 and above fulfill this request, since we need grade's resource `min` value to be at least as big as the requested value. For example, nodes in grade 2 with less than 3C and 20Gi don't fulfill our requirements, so we eliminate the entire grade due to that. Scheduler will then choose the cluster which can fit the most replicas on its nodes, which in our case is cluster 3, as computed below: | |||
There was a problem hiding this comment.
Scheduler will then choose the cluster which can fit the most replicas on its nodes, which in our case is cluster 3, as computed below
Maybe we don't need to say the intention of which clusters it will choose, because it depends on the scheduling policy, we just say the result of available replicas each cluster can hold.
There was a problem hiding this comment.
Ok, I can change that, scheduling result was mentioned in the original english text, so I included it too.
There was a problem hiding this comment.
Something like:
The Cluster Resource Model classifies nodes within a cluster based on their available resources. This classification helps the scheduler determine the number of replicas that can fit into each cluster. By categorizing nodes according to their available resources, the Cluster Resource Model enables the scheduler to make decisions about where to place application replicas based on different scheduling strategies.
[edit]: Please ignore this, I should comment it on above. Because it's a general introduction.
|
|
||
|
|
||
| Suppose now that the Pod requires 3C and 60Gi. Grade 2 nodes do not satisfy every resource request, so after considering the number of nodes available in each cluster, the scheduler prefers to schedule the Pod to member1. | ||
| Suppose now that the Pod requires 5C and 60Gi. In this case, not even grade 3 nodes satisfy the resource request (some may do, but since we can't know for sure, the entire grade has to be eliminated) since 5C > 4C and 60Gi > 32Gi, so after computing which cluster can fit the most replicas on its nodes, we get cluster 3 again, since it can fit 4 replicas, while the rest can fit none: |
There was a problem hiding this comment.
so after computing which cluster can fit the most replicas on its nodes, we get cluster 3 again, since it can fit 4 replicas, while the rest can fit none:
ditto.
Signed-off-by: LavredisG <lavredisgoume@gmail.com>
|
[APPROVALNOTIFIER] This PR is APPROVED This pull-request has been approved by: RainbowMango The full list of commands accepted by this bot can be found here. The pull request process is described here
Needs approval from an approver in each of these files:
Approvers can indicate their approval by writing |
karmada-bot
added
the
approved
/kind documentation