I'm attempting to set up a Rancher / Kubernetes dev lab on a set of four local virtual machines, however when attempting to add nodes to the cluster Rancher seems to be permanently stuck at 'Waiting to register with Kubernetes'.
From extensive Googling I suspect there is some kind of a communications problem between the Rancher node and the other three, however I can't find how to attempt to diagnose it, the instructions for finding logs in Rancher 1.x don't apply for 2.x and all the information I've so far found for 2.x appears to be on how to configure logging for a working cluster, as opposed to where to find Rancher's own logs of it's attempts to set up clusters.
So effectively two questions:
- What is the best way to go about diagnosing this problem?
- Where can I find Rancher's logs of it's cluster-building activities?
Details of my setup: Four identical VMs, all with Ubuntu 20.04 and Docker 20.10.5, all running under Proxmox on the same host and all can ping and ssh to each other. All have full Internet access.
Rancher 2.5.7 is installed on 192.168.0.180 with the other three nodes being 181-183.
Using "Global > Cluster > Add Cluster" I created a new cluster, using the default settings.
Rancher gives me the following code to execute on the nodes, this has been done, with no errors reported:
sudo docker run -d --privileged --restart=unless-stopped --net=host -v /etc/kubernetes:/etc/kubernetes -v /var/run:/var/run rancher/rancher-agent:v2.5.7 --server https://192.168.0.180 --token (token) --ca-checksum (checksum) --etcd --controlplane --worker
According to the Rancher setup instructions Rancher should now configure and take control of the nodes, however nothing happens and the nodes continue to show "Waiting to register with Kubernetes".
I've execed into the Rancher container on .180 "docker exec -it (container-id) bash" and searched for the logs, however the /var/lib/cattle directory where in older versions the debug logs were found, is empty.
Update 2021-06-23 Having got nowhere with this I deleted the existing cluster attempt in Rancher, stopped all existing Docker processes on the nodes, and tried to create a new cluster, this time using one node each for etcd, controlplane, and worker, instead of all three doing all three tasks.
Exactly the same thing happens, Rancher just forever says "Waiting to register with Kubernetes." Looking at the logs on node-1 (181) using docker ps to find the id and then docker logs to view them, I get this:
root@knode-1:~# docker ps
CONTAINER ID IMAGE COMMAND CREATED STATUS PORTS NAMES
3ca92e0ea581 rancher/rancher-agent:v2.5.7 "run.sh --server htt…" About a minute ago Up About a minute epic_goldberg
root@knode-1:~# docker logs 3ca92e0ea581
INFO: Arguments: --server https://192.168.0.180 --token REDACTED --ca-checksum 151f030e78c10cf8e2dad63679f6d07c166d2da25b979407a606dc195d08855e --etcd
INFO: Environment: CATTLE_ADDRESS=192.168.0.181 CATTLE_INTERNAL_ADDRESS= CATTLE_NODE_NAME=knode-1 CATTLE_ROLE=,etcd CATTLE_SERVER=https://192.168.0.180 CATTLE_TOKEN=REDACTED
INFO: Using resolv.conf: nameserver 127.0.0.53 options edns0
WARN: Loopback address found in /etc/resolv.conf, please refer to the documentation how to configure your cluster to resolve DNS properly
INFO: https://192.168.0.180/ping is accessible
INFO: Value from https://192.168.0.180/v3/settings/cacerts is an x509 certificate
time="2021-06-23T09:46:36Z" level=info msg="Listening on /tmp/log.sock"
time="2021-06-23T09:46:36Z" level=info msg="Rancher agent version v2.5.7 is starting"
time="2021-06-23T09:46:36Z" level=info msg="Option customConfig=map[address:192.168.0.181 internalAddress: label:map[] roles:[etcd] taints:[]]"
time="2021-06-23T09:46:36Z" level=info msg="Option etcd=true"
time="2021-06-23T09:46:36Z" level=info msg="Option controlPlane=false"
time="2021-06-23T09:46:36Z" level=info msg="Option worker=false"
time="2021-06-23T09:46:36Z" level=info msg="Option requestedHostname=knode-1"
time="2021-06-23T09:46:36Z" level=info msg="Connecting to wss://192.168.0.180/v3/connect/register with token rbdbrk8r7ncbvb9ktw9w669tj7q9xppb9scwxp9wj8zj25nhfq24s9"
time="2021-06-23T09:46:36Z" level=info msg="Connecting to proxy" url="wss://192.168.0.180/v3/connect/register"
time="2021-06-23T09:46:36Z" level=info msg="Waiting for node to register. Either cluster is not ready for registering or etcd and controlplane node have to be registered first"
time="2021-06-23T09:46:38Z" level=info msg="Starting plan monitor, checking every 15 seconds"
The only error showing appears to be the DNS one - I originally set the node's resolv.conf to use 1.1.1.1 and 8.8.4.4, so presumably the Docker install changed it, however testing 127.0.0.53 on a range of domains and records it resolves DNS correctly so I don't think that's the problem.
Help?
