search

mayadata-io / oep-e2e

Plan tests and run them for OpenEBS Enterprise Platform
Apache License 2.0
6 stars 10 forks source link

Not able to figure cluster information from the names #711

Open AmitKumarDas opened 4 years ago

AmitKumarDas commented 4 years ago 
*** System restart required ***
Last login: Mon Jun 22 10:50:52 2020 from 49.37.202.34
oep@rack2:~$ 
oep@rack2:~$ kubectl get no
NAME              STATUS   ROLES               AGE   VERSION
rancher3-master   Ready    controlplane,etcd   84d   v1.15.9
rancher3-node1    Ready    worker              84d   v1.15.9
rancher3-node2    Ready    worker              84d   v1.15.9
rancher3-node3    Ready    worker              84d   v1.15.9
oep@rack2:~$ 
oep@rack2:~$ export KUBECONFIG=~/.kube/cluster2_config
oep@rack2:~$ kubectl get no
NAME                 STATUS   ROLES               AGE   VERSION
oep-rancher-master   Ready    controlplane,etcd   65d   v1.15.11
oep-rancher-node1    Ready    worker              65d   v1.15.11
oep-rancher-node2    Ready    worker              65d   v1.15.11
oep-rancher-node3    Ready    worker              65d   v1.15.11
oep-rancher-node4    Ready    worker              65d   v1.15.11
oep-rancher-node5    Ready    worker              65d   v1.15.11
oep@rack2:~$ 
oep@rack2:~$ export KUBECONFIG=~/.kube/cluster3_config
oep@rack2:~$ 
oep@rack2:~$ 
oep@rack2:~$ kubectl get no
NAME                 STATUS   ROLES               AGE   VERSION
rancher-c3-master    Ready    controlplane,etcd   40d   v1.15.11
rancher-c3-worker1   Ready    worker              40d   v1.15.11
rancher-c3-worker2   Ready    worker              40d   v1.15.11
rancher-c3-worker3   Ready    worker              40d   v1.15.11
rancher-c3-worker4   Ready    worker              40d   v1.15.11
rancher-c3-worker5   Ready    worker              40d   v1.15.11
AmitKumarDas commented 4 years ago

1/ Can we use following naming conventions for nodes?

# c1 - indicates cluster 1
# pri - indicates primary
# sec - indicates secondary
oep-rancher-c1-pri-1 
oep-rancher-c1-sec-1
oep-rancher-c1-sec-2
oep-rancher-c1-sec-3

2/ All TCID scripts etc must show the cluster it is being run on? 3/ UI should show TCID & cluster name

AmitKumarDas commented 4 years ago

We can depend on the declarative mode of installing Kubernetes clusters. For example, terraform, cluster-api, k3sup, fluxcd, etc. can have descriptive node name that has platform details. When we run the jobs, they can kubectl get node before executing the test case.