Permalink
Browse files

Merge pull request #985 from resouer/fix-taint-key

Automatic merge from submit-queue.

Use new taint key in design doc

Part of: kubernetes/kubernetes#51246

~~**DO NOT MERGE** until kubernetes/kubernetes#51266 is in!~~
  • Loading branch information...
k8s-merge-robot committed Nov 21, 2017
2 parents 1d86f60 + c2cc0d5 commit 0b054beeee8625117995ab03e0c6247a2bbfe5d5
@@ -18,20 +18,20 @@ In addition to this, with taint-based-eviction, the Node Controller already tain
| ConditionType | Condition Status |Effect | Key |
| ------------------ | ------------------ | ------------ | -------- |
|Ready |True | - | |
| |False | NoExecute | node.kubernetes.io/notReady |
| |False | NoExecute | node.kubernetes.io/not-ready |
| |Unknown | NoExecute | node.kubernetes.io/unreachable |
|OutOfDisk |True | NoSchedule | node.kubernetes.io/outOfDisk |
|OutOfDisk |True | NoSchedule | node.kubernetes.io/out-of-disk |
| |False | - | |
| |Unknown | - | |
|MemoryPressure |True | NoSchedule | node.kubernetes.io/memoryPressure |
|MemoryPressure |True | NoSchedule | node.kubernetes.io/memory-pressure |
| |False | - | |
| |Unknown | - | |
|DiskPressure |True | NoSchedule | node.kubernetes.io/diskPressure |
|DiskPressure |True | NoSchedule | node.kubernetes.io/disk-pressure |
| |False | - | |
| |Unknown | - | |
|NetworkUnavailable |True | NoSchedule | node.kubernetes.io/networkUnavailable |
|NetworkUnavailable |True | NoSchedule | node.kubernetes.io/network-unavailable |
| |False | - | |
| |Unknown | - | |
For example, if a CNI network is not detected on the node (e.g. a network is unavailable), the Node Controller will taint the node with `node.kubernetes.io/networkUnavailable=:NoSchedule`. This will then allow users to add a toleration to their `PodSpec`, ensuring that the pod can be scheduled to this node if necessary. If the kubelet did not update the node’s status after a grace period, the Node Controller will only taint the node with `node.kubernetes.io/unreachable`; it will not taint the node with any unknown condition.
For example, if a CNI network is not detected on the node (e.g. a network is unavailable), the Node Controller will taint the node with `node.kubernetes.io/network-unavailable=:NoSchedule`. This will then allow users to add a toleration to their `PodSpec`, ensuring that the pod can be scheduled to this node if necessary. If the kubelet did not update the node’s status after a grace period, the Node Controller will only taint the node with `node.kubernetes.io/unreachable`; it will not taint the node with any unknown condition.
@@ -332,7 +332,7 @@ The term `Partitions` are used here to describe the main use cases for local sto
spec:
<snip>
nodeTolerations:
- key: node.alpha.kubernetes.io/notReady
- key: node.alpha.kubernetes.io/not-ready
operator: TolerationOpExists
tolerationSeconds: 600
- key: node.alpha.kubernetes.io/unreachable

0 comments on commit 0b054be

Please sign in to comment.