whentimeslows/website
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

Update eviction strategy to include priority (#6949)

Browse Source
This commit is contained in:
David Ashpole
2018-01-25 14:21:39 -08:00
committed by k8s-ci-robot
parent e1dbc1a497
commit de7b93abb7
1 changed files with 14 additions and 17 deletions
Download Patch File Download Diff File Expand all files Collapse all files
docs/tasks/administer-cluster/out-of-resource.md
+14 -17
View File
@@ -196,25 +196,22 @@ If `nodefs` filesystem has met eviction thresholds, `kubelet` frees up disk spac
If the `kubelet` is unable to reclaim sufficient resource on the node, `kubelet` begins evicting Pods. If the `kubelet` is unable to reclaim sufficient resource on the node, `kubelet` begins evicting Pods.
The `kubelet` ranks Pods for eviction first by their quality of service, and then by the consumption The `kubelet` ranks Pods for eviction first by whether or not their usage of the starved resource exceeds requests,
of the starved compute resource relative to the Pods' scheduling requests. then by [Priority](https://kubernetes.io/docs/concepts/configuration/pod-priority-preemption/), and then by the consumption of the starved compute resource relative to the Pods' scheduling requests.
As a result, `kubectl` ranks and evicts Pods in the following order: As a result, `kubelet` ranks and evicts Pods in the following order:
* `BestEffort` Pods consume the most of the starved resource are failed first. * `BestEffort` or `Burstable` Pods whose usage of a starved resource exceeds its request.
Local disk is a `BestEffort` resource. Such pods are ranked by Priority, and then usage above request.
* `Burstable` Pods consume the greatest amount of the starved resource * `Guaranteed` pods and `Burstable` pods whose usage is beneath requests are evicted last.
relative to their request for that resource are killed first. If no Pod `Guaranteed` Pods are guaranteed only when requests and limits are specified for all
has exceeded its request, the strategy targets the largest consumer of the the containers and they are equal. Such pods are guaranteed to never be evicted because
starved resource. of another Pod's resource consumption. If a system daemon (such as `kubelet`, `docker`,
* `Guaranteed` Pods are guaranteed only when requests and limits are specified and `journald`) is consuming more resources than were reserved via `system-reserved` or
for all the containers and they are equal. A `Guaranteed` Pod is guaranteed to `kube-reserved` allocations, and the node only has `Guaranteed` or `Burstable` Pods using
never be evicted because of another Pod's resource consumption. If a system less than requests remaining, then the node must choose to evict such a Pod in order to
daemon (such as `kubelet`, `docker`, and `journald`) is consuming more resources preserve node stability and to limit the impact of the unexpected consumption to other Pods.
than were reserved via `system-reserved` or `kube-reserved` allocations, and the In this case, it will choose to evict pods of Lowest Priority first.
node only has `Guaranteed` Pods remaining, then the node must choose to evict a
`Guaranteed` Pod in order to preserve node stability and to limit the impact
of the unexpected consumption to other `Guaranteed` Pods.
If necessary, `kubelet` evicts Pods one at a time to reclaim disk when `DiskPressure` If necessary, `kubelet` evicts Pods one at a time to reclaim disk when `DiskPressure`
is encountered. If the `kubelet` is responding to `inode` starvation, it reclaims is encountered. If the `kubelet` is responding to `inode` starvation, it reclaims