Expose IPI_ROOTVOLUME_SIZE input in CLI when configuring SNO clusters in AWS

[alequint]

Feature and motivation

Size of Manage image increased recently and we need to have additional ephemeral storage to complete the work. Today, it is using the default value, which should be around 100 / 100GB. Setting IPI_ROOTVOLUME_SIZE allow us to change that to a more acceptable value, around 250GB.

Automation already have this input, in ocp_provision role, but we need to promote it also to CLI

Usage example

We should be capable to run that in a non-interactive and interactive mode as all the settings. During interactive mode, if SNO is detected, the value must be enforced to 250GB (?)

References

Internal slack thread here

[alequint]

@unnati-solanki-git as we discussed, considering provisioning command runs ansible-devops playbook and considering that playbook/role is already prepared to receive IPI_ROOTVOLUME_SIZE, for this particular issue we have to enforce the IPI_ROOTVOLUME_SIZE=250 (following documentation requirement) in case an SNO environment is detected (i.e. when the number of nodes being asked to deploy is one)

[kthomsonAQ]

Please consider NOT enforcing the 250GB limit.

[alequint]

@kthomsonAQ the main issue reported is that Manage image really increased its size and the existent default ephemeral storage used today is not accommodating that. So 250 would be the minimum acceptable at the moment for building Manage image, otherwise it may not even be installed moving forward.

@kthomsonAQ would you mind to share more details about your concern with ephemeral storage requirement?

[mudspringhiker]

Performance team recommends to have 300GB.

[kthomsonAQ]

This technote has been published. https://www.ibm.com/support/pages/node/6998735

We can work around this issue, and I have been able to deploy a cluster in AWS. However I feel that the default should be set to 300GB and it would be really helpful to expose the IPI_ROOTVOLUME_SIZE as a parameter that is required, (defaulting to 300GB)