You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
I'm currently reviewing the example tfvars files of the 'ha' and 'minimal' deployment. I notice that the instance types r5 and m5 are used, rather than r5d or m5d.
As far as I am aware, the distinction between these instance types is that the r5 and m5 type do not come with any 'instance (ephemeral) storage', yet the r5d and m5d types do.
Is it the intention of these examples that ephemeral disk is available not used for SASWORK and CASCACHE (minimal example) and SASWORK (ha example)?
Also, will in the minimal example, should the generic node have the labels and taints for the stateful and stateless services?
Expected Behavior
e.g. should the compute nodes use m5d.8xlarge instances rather than m5.8xlarge as is provided in the ha example tfvars file?
The 'minimal' example uses:
CAS: r5.xlarge (no ephemeral disk included with this instance type)
Compute/Generic: m5.2xlarge (no ephemeral disk included with this instance type)
The 'ha' example uses:
CAS: i3.8xlarge (includes 4 x 1900[GB] NVMe SSD of emphermal disk)
Compute: m5.8xlarge (no ephemeral disk included with this instance type)
Terraform Version Details
No response
Terraform Variable File Details
No response
Steps to Reproduce
I'm currently reviewing the example tfvars files of the 'ha' and 'minimal' deployment. I notice that the instance types r5 and m5 are used, rather than r5d or m5d.
As far as I am aware, the distinction between these instance types is that the r5 and m5 type do not come with any 'instance (ephemeral) storage', yet the r5d and m5d types do.
Is it the intention of these examples that ephemeral disk is available not used for SASWORK and CASCACHE (minimal example) and SASWORK (ha example)?
Also, will in the minimal example, should the generic node have the labels and taints for the stateful and stateless services?
Expected Behavior
e.g. should the compute nodes use m5d.8xlarge instances rather than m5.8xlarge as is provided in the ha example tfvars file?
Actual Behavior
Additional Context
No response
References
https://aws.amazon.com/ec2/instance-types/
Code of Conduct
The text was updated successfully, but these errors were encountered: