error execution phase upload-config/kubelet: Error writing Crisocket information for the control-plane node: timed out waiting for the condition #1438
Comments
neolit123
added
help wanted
|
Hi, @timothysc, I see that you added this to a milestone for v1.15, am I hitting a known problem? |
|
@hreidar i don't believe we have gotten a similar report about external ETCD HA clusters not being able to write CRI socket. also, i can't seem to find a problem with your setup. what happens if you try to
anything else that is interesting in the kubelet logs? |
|
@neolit123 thanks for the info. I agree. I don't see anything wrong and after the error I'm able to query the kubernetes API and I can see k8s data in my etcd cluster. I'll try to make some more tests tomorrow. I will report if I find something interesting. |
|
Can you share the logs of your kube-apiserver container @hreidar |
|
If I'm not worng, the drop in 20-etcd-service-manager.conf should not exists on control-plane nodes
|
|
@codeaholic I can but not until Tuesday as I'm at home with my two girls that are ill at the moment. @fabriziopandini there is a possibility that I have generated this file by accident as I was preparing the nodes all at once. Does this influence the other k8s components in some way? etcd should not installed on the master in my case. |
|
Hi, @codeaholic, @fabriziopandini, I had to redo my test cluster as my setup was fubar when I got to work on Tuesday. This time around I got things to work it seems. The only thing I did differently this time is I ran kubeadm commands using sudo instead of root. Is this something that is known to screw things up or .... ? If so it should probably be documented somewhere. |
|
Got the same issue on my local setup |
drop 20-etcd-service-manager.conf, it works! |
|
I solved this issue. docker info | grep -i cgroup /var/lib/kubelet/config.yaml Because of this, I can run "kubelet". So I tried to match for this value. In /etc/systemd/system/kubelet.service.d/ In 10-kubeadm.conf
|
|
Got the same issue on my external setup: |
|
I use kubeadm v1.15.0 on Ubuntu 16.04.2 LTS and same issue when running |
|
I got this when I had ran init, then manually cleared down the machine, and then ran it again, while testing etc. Running |
Right command is sudo kubeadm reset, and follow its output at least to delete /etc/cni/net.d and $HOME/.kube/config. |
|
[kubelet-check] Initial timeout of 40s passed. |
|
How to resloved this issue |
too few information |
|
try the support channels: if a client cannot write information to a Node object (such as the CRI socket information) due to timeout, then something is not OK with the api-server in the cluster. |
|
I have the same error but I don't know what wrong with my steps I plan to setup with below boxs Install docker, kubelet kubeadm kubectl on all cluster node HAProxy setup backend k8s_backend Keepalived setup root@haproxy01:~# cat /etc/keepalived/keepalived.conf vrrp_instance VI_1 { Generating the TLS certificates $ vim ca-config.json $ vim ca-csr.json $ cfssl gencert -initca ca-csr.json | cfssljson -bare ca cfssl gencert scp ca.pem kubernetes.pem kubernetes-key.pem [email protected]:/etc/etcd/ Etcd on two masters setup [Unit] [Service] [Install] root@master02:~# cat /etc/systemd/system/etcd.service [Service] [Install] root@master01:~# ETCDCTL_API=3 etcdctl member list Prepare cluster config file kind: KubeletConfiguration root@master01:~# kubeadm init --config=cluster.yaml |
|
Your setup seems fine on a surface look.
You might get better analysis at the support channels. See the support
section if the readme.
error execution phase upload-config/kubelet: Error writing Crisocket
information for the control-plane node: timed out waiting for the condition
This means kubeadm as a client is timing out accesing an apiserver behind a
load balancer. So there must be a blackout at the LB for some reason at
that time.
|
|
The load balancer forwards the request to kube-apiserver in two masters How can I verify the kube-apiserver properly or not |
|
i tried all solutions but i have the same error everytime , Any help |
check kubelet log may help. BTW, this seems to be a support case from you and we suggest to use slack or StackOverflow to do that. |
|
I believe my reason was becuase i used the --node-name $NODENAME which was adding my host name and and my hostname was uppercase which i think it may of not supported. |
Hi, I'm trying out kubeadm and I'm using the official docs in setting up a HA cluster. I have managed to create a etcd cluster but the init step on the first master node is failing with the following error:
What keywords did you search in kubeadm issues before filing this one?
error execution phase upload-config/kubelet
and got:
#1382
#1227
Is this a BUG REPORT or FEATURE REQUEST?
BUG REPORT
Versions
Note: Also tryed Docker version 17.03.3-ce and 18.09.3-ce
I'm running this on vm's in vmware.
What happened?
What you expected to happen?
Master to initialize without issue.
How to reproduce it (as minimally and precisely as possible)?
node preperation
config creation for kublet and etcd
generate and distribute certs
config and init master nodes
###Anything else we need to know?
cluster nodes
k8s-c3-lb - 10.10.10.76
k8s-c3-e1 - 10.10.10.90
k8s-c3-e2 - 10.10.10.91
k8s-c3-e3 - 10.10.10.92
k8s-c3-m1 - 10.10.10.93
k8s-c3-m2 - 10.10.10.94
k8s-c3-m3 - 10.10.10.95
k8s-c3-w1 - 10.10.10.96
k8s-c3-w2 - 10.10.10.97
k8s-c3-w3 - 10.10.10.98
nginx LB config
netcat ouput to check lb
kubeadm config on etcd nodes
etcd check from master
kubeadm config on master
docker ps
journalctl -xeu kubelet
systemctl
The text was updated successfully, but these errors were encountered: