You are viewing a plain text version of this content. The canonical link for it is here.
Posted to issues@yunikorn.apache.org by "Peter Bacsko (Jira)" <ji...@apache.org> on 2023/05/17 18:04:00 UTC
[jira] [Comment Edited] (YUNIKORN-1747) Improve the performance of nodeInfoListerImpl.HavePodsWithAffinityList()
[ https://issues.apache.org/jira/browse/YUNIKORN-1747?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17723558#comment-17723558 ]
Peter Bacsko edited comment on YUNIKORN-1747 at 5/17/23 6:03 PM:
-----------------------------------------------------------------
alloc_space sample from one of our upstream user:
!image-2023-05-17-20-02-57-177.png!
was (Author: pbacsko):
alloc_space sample from our upstream user:
!image-2023-05-17-20-02-57-177.png!
> Improve the performance of nodeInfoListerImpl.HavePodsWithAffinityList()
> ------------------------------------------------------------------------
>
> Key: YUNIKORN-1747
> URL: https://issues.apache.org/jira/browse/YUNIKORN-1747
> Project: Apache YuniKorn
> Issue Type: Sub-task
> Components: shim - kubernetes
> Reporter: Peter Bacsko
> Assignee: Peter Bacsko
> Priority: Major
> Attachments: image-2023-05-17-20-02-57-177.png
>
>
> Similarly to YUNIKORN-1746, the method {{nodeInfoListerImpl.HavePodsWithAffinityList()}} is called very often, for every pod.
> {noformat}
> func (n nodeInfoListerImpl) HavePodsWithAffinityList() ([]*framework.NodeInfo, error) {
> nodes := n.cache.GetNodesInfoMap()
> result := make([]*framework.NodeInfo, 0, len(nodes))
> for _, node := range nodes {
> if len(node.PodsWithAffinity) > 0 {
> result = append(result, node)
> }
> }
> return result, nil
> }
> {noformat}
> This is more tricky though, but still doable. We need to know whether we should include a node to our "result" slice or not. If a node info changes, a counter value changes with it. It is maintained inside {{{}framework.NodeInfo{}}}:
> {noformat}
> // NodeInfo is node level aggregated information.
> type NodeInfo struct {
> // Overall node information.
> node *v1.Node
> // Pods running on the node.
> Pods []*PodInfo
> ...
> // Whenever NodeInfo changes, generation is bumped.
> // This is used to avoid cloning it if the object didn't change.
> Generation int64
> {noformat}
> The algorithm can go as follows:
> * First, we build a slice and save it
> * Save the per-node generation value to a map
> * Upon subsequent calls, we retrieve {{n.cache.GetNodesInfoMap()}}
> * We check if the latest {{node.Generation}} is different from our saved value
> * If it is, just rebuild the entire "result" slice, save it and update the per-node generation data
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
---------------------------------------------------------------------
To unsubscribe, e-mail: issues-unsubscribe@yunikorn.apache.org
For additional commands, e-mail: issues-help@yunikorn.apache.org