Is this a new feature, an enhancement, or a change to existing functionality?
Enhancement
How would you describe the priority of this feature request
Critical (currently preventing usage)
Please provide a clear description of problem this feature solves
Booting the Scout image and running it at steady-state consumes roughly 10GB of RAM. This is a blocker when trying to run Scout in resource constrained environments such as in a VM.
I am running the x86_64 Scout image.
Feature Description
As a user/tester of Scout, I would like the memory consumption of Scout at boot time and steady-state to be greatly reduced, so that I can run Scout in resource-constrained environments.
Describe your ideal solution
Other OS images such as the Ubuntu cloud image are designed to run with extremely low resource requirements, often 1 vCPU and 1GB RAM (Ubuntu Server docs). Since Scout doesn't need to support all the functionality of a general-purpose OS, I would expect that it can run in similar situations.
Describe any alternatives you have considered
I have considered short-circuiting the NICO lifecycle state machine and skipping the Scout step/callback altogether, although I'd like to stay as close to upstream NICO as possible.
Additional context
scout_os_memory.log
Please find attached a log at 1s-intervals of a sample VM booting Scout. You can clearly see the RSS/PSS sizes balloon to ~10GB as Scout is booting.
Code of Conduct
Is this a new feature, an enhancement, or a change to existing functionality?
Enhancement
How would you describe the priority of this feature request
Critical (currently preventing usage)
Please provide a clear description of problem this feature solves
Booting the Scout image and running it at steady-state consumes roughly 10GB of RAM. This is a blocker when trying to run Scout in resource constrained environments such as in a VM.
I am running the x86_64 Scout image.
Feature Description
As a user/tester of Scout, I would like the memory consumption of Scout at boot time and steady-state to be greatly reduced, so that I can run Scout in resource-constrained environments.
Describe your ideal solution
Other OS images such as the Ubuntu cloud image are designed to run with extremely low resource requirements, often 1 vCPU and 1GB RAM (Ubuntu Server docs). Since Scout doesn't need to support all the functionality of a general-purpose OS, I would expect that it can run in similar situations.
Describe any alternatives you have considered
I have considered short-circuiting the NICO lifecycle state machine and skipping the Scout step/callback altogether, although I'd like to stay as close to upstream NICO as possible.
Additional context
scout_os_memory.log
Please find attached a log at 1s-intervals of a sample VM booting Scout. You can clearly see the RSS/PSS sizes balloon to ~10GB as Scout is booting.
Code of Conduct