Closed vikaschoudhary16 closed 6 years ago
DiskPressure True Sat, 10 Mar 2018 05:33:46 +0000 Sat, 10 Mar 2018 02:39:48 +0000 KubeletHasDiskPressure kubelet has disk pressure
I0310 06:32:39.985159 55790 helpers.go:829] eviction manager: observations: signal=allocatableMemory.available, available: 131710276Ki, capacity: 131815376Ki
I0310 06:32:39.985171 55790 helpers.go:843] eviction manager: thresholds - ignoring grace period: threshold [signal=imagefs.available, quantity=8049131839] observed 7809724Ki
I0310 06:32:39.985179 55790 helpers.go:843] eviction manager: thresholds - reclaim not satisfied: threshold [signal=imagefs.available, quantity=8049131839] observed 7809724Ki
I0310 06:32:39.985193 55790 eviction_manager.go:284] eviction manager: node conditions - observed: [DiskPressure]
After freeing up some space of root file-system, it got completed without any problems.
Answer to the question that why it was working with 3.7 and failing with 3.9 only is because following code block which is there is 3.9 and not in 3.7, in fact whole webconsole.go :
err = wait.Poll(1*time.Second, 10*time.Minute, func() (bool, error) {
glog.V(2).Infof("polling for web console server availability")
ds, err := kubeClient.Extensions().Deployments(consoleNamespace).Get("webconsole", metav1.GetOptions{})
if err != nil {
return false, err
}
if ds.Status.ReadyReplicas > 0 {
return true, nil
}
return false, nil
})
Closing the issue.
On the same machine, if I switch to release-3.7 it works.
Version
oc v3.9.0-alpha.4+9fd063a-531 kubernetes v1.9.1+a0ce1bc657 features: Basic-Auth GSSAPI Kerberos SPNEGO
Steps To Reproduce
Current Result
oc cluster up
fails after hanging for 10 minutesExpected Result
No errors
Additional Information