vmware-tanzu / community-edition

VMware Tanzu Community Edition is no longer an actively maintained project. Code is available for historical purposes only.
https://tanzucommunityedition.io/
Apache License 2.0
1.34k stars 308 forks source link

Capture time to guest cluster for TCE #133

Closed joshrosso closed 3 years ago

joshrosso commented 3 years ago

We'd like to be able to get a human up and running with a cluster that can run workloads in 10 minutes or less.

Using the v0.1.0 release, essentially tkg 1.2.1, time a cluster bootstrapping in AWS to get our current metric.

Start timing after downloading tkg and stop timing after the guest cluster nodes report Ready.

seemiller commented 3 years ago

Using the below steps, I created management and worker clusters.

The management cluster took 11 minutes 14 seconds to spin up. The worker cluster took 7 minutes 53 seconds to spin up.

So, roughly 19 minutes to get a usable cluster.


Steps to create a TKG Cluster, AWS Edition

After downloading, un-tar and get the executable on the path

tar xvf tkg-linux-amd64-v1.2.1-vmware.1.tar
sudo cp -p tkg/tkg-linux-amd64-v1.2.1+vmware.1 /usr/local/bin/tkg

Export your AWS credentials

export AWS_ACCESS_KEY_ID=ABCDEFGHIJKLMNOPQRST
export AWS_SECRET_ACCESS_KEY=ABC12+ABCDEFGHIJKLMNOPQRSTUVWXYZ01234567

Set the AWS Region

export AWS_REGION=us-east-1

Set the name of an AWS SSH key-pair

export AWS_SSH_KEY_NAME=default

Set a name for the management cluster

export MANAGEMENT_CLUSTERNAME=manager

Create the management cluster

tkg init --infrastructure=aws --name=${MANAGEMENT_CLUSTERNAME} --plan=dev

Set a name for the worker cluster

export WORKER_CLUSTERNAME=worker

Create the worker cluster

tkg create cluster ${WORKER_CLUSTERNAME} --plan=dev

Enjoy your cluster

tkg get credentials ${WORKER_CLUSTERNAME}
kubectl apply -f ...

Or just delete them!

tkg delete cluster ${WORKER_CLUSTERNAME}

# Have to wait for the worker to be deleted before you can delete the management cluster
# This takes awhile as well
tkg delete management-cluster ${MANAGEMENT_CLUSTERNAME}
stmcginnis commented 3 years ago

Just to get this captured somewhere, here's some data on deployment to a vSphere cluster. Local single node lab environment.

Hypervisor: VMware ESXi, 7.0.1, 17325551
Model:  System Product Name
Processor Type: AMD Ryzen 7 1700X Eight-Core Processor
Logical Processors: 16
NICs:   2

Cloning to-from the same 1TB drive.

Time not including setup of tkg CLI.

Management cluster deployment: ~11 minutes Workload cluster deployment: 5:37