Host Preparation After Cluster Extension Configuration
With AWS cloud resource access configured, the next step is to modify the device entries created using the cmcluster-extension to leverage the correct categories and make changes when necessary to public cloud settings.
Rename the nodes according to their expected usage as follows from the device submenu.
1cmsh 2device 3rename us-west-2-cnode001 us-west-2-knode001 4rename us-west-2-cnode002 us-west-2-knode002 5rename us-west-2-cnode003 us-west-2-knode003 6rename us-west-2-cnode004 us-west-2-gpu-node001 7commit
Update the categories for all four public cloud nodes.
1set us-west-2-knode001 category k8s-cloud-master 2set us-west-2-knode002 category k8s-cloud-master 3set us-west-2-knode003 category k8s-cloud-master 4set us-west-2-gpu-node001 category k8s-cloud-gpu-worker 5commit
Increase the EBS volume size to 100 GiB on the knode systems.
1cloudsettings us-west-2-knode001 2storage 3set ebs size 100GiB 4commit 5.. 6.. 7.. 8cloudsettings us-west-2-knode002 9storage 10set ebs size 100GiB 11commit 12.. 13.. 14.. 15cloudsettings us-west-2-knode003 16storage 17set ebs size 100GiB 18commit 19.. 20.. 21..
Increase the EBS volume size of us-west-2-gpu-node001 to 100 GiB and change the instancetype for us-west-2-gpu-node001 to g4dn.xlarge (a lower-cost GPU instance with a single T4 GPU).
If requirements justify a higher instance spec, use the appropriate instance type—any NVIDIA GPU instance should work.
1cloudsettings us-west-2-gpu-node001 2set instancetype g4dn.xlarge 3commit 4storage 5set ebs size 100GiB 6commit
Increase the EBS volume size to 200 GiB on the director system.
1.. 2.. 3cloudsettings us-west-2-director 4storage 5set ebs size 200GiB 6commit
Update the softwareimage for the aws-cloud-director category.
1category 2use aws-cloud-director 3set softwareimage cloud-director-image
Update disksetup for the cloud director to use the same partitioning scheme set for the other public cloud nodes.
1set disksetup /tmp/big-cloud-disk.xml 2commit