Closed ZihanJiang96 closed 9 months ago
Is this a BUG REPORT or FEATURE REQUEST?:
BUG REPORT
What happened:
time="2023-11-14T04:40:15Z" level=error msg="failed to drain node ip-10-184-115-74.us-west-2.compute.internal error: error cordoning node: cordon error: Patch \"https://172.20.0.1:443/api/v1/nodes/ip-10-184-115-74.us-west-2.compute.internal\": dial tcp 172.20.0.1:443: i/o timeout "
internally, kubectl lib mark the the node obj's Spec.Unschedulable to true
node.Spec.Unschedulable = true
Unschedulable
true
nil
What you expected to happen:
lifecycle-manager should make sure the cordon operation succeed before drain the node.
How to fix this issue:
create a copy of the node obj and pass the copy to the drainNodeUtil() func
node
Nice find!
Is this a BUG REPORT or FEATURE REQUEST?:
BUG REPORT
What happened:
internally, kubectl lib mark the the node obj's Spec.Unschedulable to true
node.Spec.Unschedulable = true
Unschedulable
already set totrue
, so it thinks there is no need to cordon this node and returnnil
directly.What you expected to happen:
lifecycle-manager should make sure the cordon operation succeed before drain the node.
How to fix this issue:
create a copy of the
node
obj and pass the copy to the drainNodeUtil() func