sched: evaluate error message once #111295
Conversation
k8s-ci-robot
added
release-note-none
|
@Huang-Wei: This issue is currently awaiting triage. If a SIG or subproject determines this is a relevant issue, they will accept it by applying the The Instructions for interacting with me using PR comments are available here. If you have questions or suggestions related to my behavior, please file an issue against the kubernetes/test-infra repository. |
k8s-ci-robot
added
the
needs-priority
|
[APPROVALNOTIFIER] This PR is APPROVED This pull-request has been approved by: Huang-Wei The full list of commands accepted by this bot can be found here. The pull request process is described here
Needs approval from an approver in each of these files:
Approvers can indicate their approval by writing |
k8s-ci-robot
added
the
approved
|
/retest |
|
test error refer to #111012 |
| if err == ErrNoNodesAvailable { | ||
| klog.V(2).InfoS("Unable to schedule pod; no nodes are registered to the cluster; waiting", "pod", klog.KObj(pod)) | ||
| } else if fitError, ok := err.(*framework.FitError); ok { | ||
| // Inject UnschedulablePlugins to PodInfo, which will be used later for moving Pods between queues efficiently. | ||
| podInfo.UnschedulablePlugins = fitError.Diagnosis.UnschedulablePlugins | ||
| klog.V(2).InfoS("Unable to schedule pod; no fit; waiting", "pod", klog.KObj(pod), "err", err) | ||
| klog.V(2).InfoS("Unable to schedule pod; no fit; waiting", "pod", klog.KObj(pod), "err", errMsg) |
There was a problem hiding this comment.
I am not sure about this, FitError has other info beside the error msg, I am afraid we will lost it after this change.
There was a problem hiding this comment.
Fine, FitError.Error() will do that for us.
/lgtm
k8s-ci-robot
added
the
lgtm
|
The Kubernetes project has merge-blocking tests that are currently too flaky to consistently pass. This bot retests PRs for certain kubernetes repos according to the following rules:
You can:
/retest |
What type of PR is this?
/kind cleanup
/sig scheduling
What this PR does / why we need it:
Evaluating FitError is not trivial, esp. in a large cluster, as under the hood it runs sorting and aggregating the individual errors generated on each Node.
This PR leveraged the pre-work of #111036, so that we can unify the callers in one function (
handleSchedulingFailure()) and hence evaluate the error message only once.Which issue(s) this PR fixes:
Related with #110871 and #103853.
Special notes for your reviewer:
Does this PR introduce a user-facing change?