[
https://issues.apache.org/jira/browse/FLINK-24150?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17793103#comment-17793103
]
chaoran.su commented on FLINK-24150:
------------------------------------
[~asardaes] this is not the same, if we need flink cluster's pod to be a
graranteed qos level pod, we can not set limit-factor, which will make the
pod's qos level to Burstable. For flink community, there is no way to have a
resource config beyond flink configuration and still confirm the pod as a
graranteed qos level pod.
> Support to configure cpu resource request and limit in pod template
> -------------------------------------------------------------------
>
> Key: FLINK-24150
> URL: https://issues.apache.org/jira/browse/FLINK-24150
> Project: Flink
> Issue Type: New Feature
> Components: Deployment / Kubernetes
> Reporter: Yang Wang
> Priority: Major
>
> Why Flink needs to overwrite memory resource defined in pod template?
> The major reason is that Flink need to ensure the consistency between Flink
> configuration
> (\{{taskmanager.memory.process.size}} , {{kubernetes.taskmanager.cpu}}) and
> pod template resource settings. Since users could specify the total process
> memory or detailed memory[2], Flink will calculate the pod resource
> internally.
>
> For the CPU case the template’s requests/limits should have priority if they
> are specified. The factor could still be used if the template doesn’t specify
> anything. The logic could be something like this:
> # To choose CPU request
> # Read pod template first
> # If template doesn’t have anything, read from {{kubernetes.taskmanager.cpu}}
> # If configuration is not specified, fall back to default
> # To choose CPU limit
> # Read from template first
> # If template doesn’t have anything, apply factor to what was chosen in step
> 1, where the default factor is 1.
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
