minikube start hung: etcd CrashLoopBackoff #2765
Comments
|
I have the same issue except that I am using the hyperkit driver, and I'm installing a k8s 1.9.0 cluster. Gets stuck at "Starting cluster components" This is on MacOS High Sierra (10.13.3) |
|
Same issue for me. |
|
If you have an old Try starting up again. |
|
same issue here Also try But it not worked Use minikube logs -f and see the error is: Apr 24 15:44:59 minikube kubelet[3287]: E0424 15:44:59.532988 3287 reflector.go:205] k8s.io/kubernetes/pkg/kubelet/kubelet.go:451: Failed to list *v1.Service: Get https://192.168.99.100:8443/api/v1/services?limit=500&resourceVersion=0: dial tcp 192.168.99.100:8443: getsockopt: connection refused |
|
Similar problem on minikube for windows. I did a |
|
Same issue on Linux Mint (ubuntu derived distro). |
|
Same issue with me : Apr 26 21:37:10 minikube kubelet[3154]: E0426 21:37:10.657560 3154 kubelet_node_status.go:106] Unable to register node "minikube" with API server: Post https://192.168.64.22:8443/api/v1/nodes: dial tcp 192.168.64.22:8443: getsockopt: connection refused |
|
Same issue with same config as @vinceskahan also getting the |
|
Same issue here, on MacOS w/ hyperkit driver, kubeadm bootstrapper, k8s version v1.9.6. Looking into it a bit farther, it looks like Updating the k8s version to TLDR looks like Here's a reference link re: the |
|
just to share that on p.s. the |
|
update - the '--extra-config' that gsaslis mentioned above did 'not' work for me. Total hang to the point where the computer almost totally locked up and I had to reset it. Downgrading minikube to 0.25.2 'did' work just fine, so the problem appears to be something in the 0.26.x codebase. |
|
I have the same issue with hyperkit v0.20171204-60-g0e5b6b driver and MacOS High Sierra 10.13.4. In minikube logs got: Neither removing .minikube directory nor |
|
I'm pretty sure the issue is in this line, which was added as part of this commit in the lead up to the 0.26 release. AFAICT, the |
|
You can work around this by using the localkube bootstrapper. I hope a fix for this is forthcoming soon. |
|
Temporary workaround: Using these instructions as an example, I forced an install of 0.25.2 in the meantime: Using this file: |
|
The fix still isn't clear to me. What versions of Kubernetes is this reproducing with? It looks to me like the require-kubeconfig flag was only removed for 1.10, and should still have been present in 1.9: |
|
This happens on Linux (Fedora), not just mac and seems to be a problem with versions > |
|
As @dlorenc said it looks like bug caused by removing the require-kubeconfig flag. It worked perfectly fine v1.10.0 |
Which versions of Kubernetes (not minikube) are you seeing this with? |
I was seeing this when trying to run k8s v1.9.4 and v1.9.7 locally. I had to downgrade to 0.25.2 in order to make it work. |
There are known issues with the current latest versions of helm and minikube which is causing failures for new users trying to deploy zenko on minikube and therefore explicitly declaring known working versions for future reference. Issues with new versions of minikube kubernetes/minikube#2765 Issues with new versions of helm helm/helm#3991
|
same issues , failed to run |
|
|
try to run |
|
@playniuniu 不行，我测试了从0.26以上的版本都不可以，我现在在测试你说的启动方式 |
|
@playniuniu 我试了下，为什么cluster是关闭的呢？ ` ➜ ~ minikube status |
|
@Robot684 我的也是，应该不影响使用吧 |
|
@playniuniu 你可以进行后续的模拟创建pod测试吗？ |
|
@playniuniu / @Robot684 please note that when you write a comment on a GitHub issue (which has so far 25+ participants), by default, a notification email is sent to each and everyone of them. when your comment is not in english, you are essentially sending us messages in your own language, which we all can't understand. that doesn't feel very nice. please reconsider using just English. ;) |
|
same problem on macOS High Sierra 10.13.4 with minikube v0.27.0 maybe a different cause log: problem solved with localkube |
|
I too am encountering the same set of problems :- Configuration
Issues:See attempts below that did not work :- minikube start --vm-driver=hyperkit --bootstrapper=localkubeminikube start |
|
+1 |
|
Still hanging for me - mac, linux and windows - on the new 0.27 (unless the deprecated --bootstrapper=localkube is applied) |
|
@vinaynair Maybe you cannot pull the images, try to run |
|
Same here - High Sierra and had to use |
|
Similar error as @vinaynair, I had thought it was my environment issue, while it seems a common one. |
|
I think this was fixed with #2791 |
|
It does not appear that fixed the issue, because I encountered this issue two days ago and only was able to solve it by using the workaround. |
|
agree - this is 'not' fixed in 0.27.0 without using the workaround |
|
Ya, and |
|
I think there might have been several different underlying issues being tracked in here. Would someone still experiencing this mind opening another issue with all the info from the template? |
|
I think closing an issue and asking a user to create a brand new issue from a template is .... perhaps not ideal. |
|
Unfortunately there's just not enough information in any of these comments beyond the initial ones (involving the --require-kubeconfig flag, which was fixed in the linked PR). I'm sure there are other issues, but the best way to help us find and fix them is to include the debugging information asked for in the template. |
|
@dlorenc - I have to call baloney on that one. Please read 'my' original problem description for this ticket at the very top. The expectation is that I can run "minikube start" per the docs, with no options or flags or workarounds and have the blasted thing actually start. If that's somehow unclear, I can't help you. All the other comments, in my reading, are people mentioning:
It's broken if it used to work before. It's broken if deprecated workaround flags are required. So it's still broken for me (the original poster) on the same High Sierra Macbook that I reported it on. And I 'did' try to use 0.27.0 and did verify it is still broken here on that platform. |
|
I'm going to lock this issue in the interest of keeping things productive. If you'd like to help see issues get fixed, please open a new issue with the output of "minikube info" and "minikube logs" attached. Without that we can't really debug anything. |
tstromberg
added
kind/bug
tstromberg
changed the title
tstromberg
added
co/etcd
and others
BUG REPORT
Environment:
Minikube version
v0.26.1
OS
MacOS High Sierra 10.13.4
VM Driver
Virtual Box
ISO version
Latest as of today
Install tools:
Others:
What happened:
Initial minikube start hangs at 'Starting cluster components...' and never completes. Note that the vm 'is' running, and I can log into it from the Virtual Box console. From another terminal window, 'minikube ip' does return an ip address.
What you expected to happen:
Ummm - for it to complete, of course...
How to reproduce it (as minimally and precisely as possible):
minikube start
Output of
minikube logs(if applicable):(attached)
minikube-logs-output.txt
The text was updated successfully, but these errors were encountered: