5 steps to Build & Test the Edge Microvisor Toolkit #360
Replies: 3 comments 2 replies
-
@biapalmeiro: How to build the ISO image with custom version? (From target, how can i know which version of ISO was flashed?) |
Beta Was this translation helpful? Give feedback.
-
Beta Was this translation helpful? Give feedback.
-
@anujm1 @krishnajs any suggestion? We have an upcoming release with customer, and they are customizing the ISO image (dracut, cmdline, new packages, etc). And they critically want to know from the target, what image was used to burn. Any build time argument also is fine, which they can include. |
Beta Was this translation helpful? Give feedback.
Uh oh!
There was an error while loading. Please reload this page.
Uh oh!
There was an error while loading. Please reload this page.
-
1. Clone the Stable Repository
Start by cloning the Edge Microvisor Toolkit repository and checkout the recommended tag (e.g.,
3.0.x
) rather thanmain
to ensure stability:2. Install Prerequisites
Make sure you have dependencies like:
make
,gcc
,rpm
,qemu-utils
python3
,pip
, and standard build toolsThese are essential for the toolchain and packaging process.
3. Build the Toolchain
Use the included Makefile to build the customized build environment:
This sets up the compilation environment and local build tools necessary for packaging.
4. Generate an OS Image (ISO or RAW)
Select one of the predefined JSON configs such as
edge-image.json
oredge-image-rt.json
and run:OR for RAW images:
You can optionally include
REBUILD_PACKAGES=n
to use prebuilt RPMs and speed up builds.5. Flash & Test Your Image
Once the build completes:
Write the ISO to USB using
dd
or your preferred imaging tool.Boot the target edge node or VM.
Perform basic validation:
Login (default credentials may vary by image).
Run
cat /etc/os-release
to confirm version.Use
uname -r
andmount
to check kernel and partition setup.Inspect read-only rootfs and update mechanisms (A/B update support).
And 3 Whys This Matters for Edge AI Developers
Full customization: Tailor low-footprint OS images optimized for your inference workloads.
Immutable, secure base: Leverage EMT's A/B updates, dm‑verity, Secure Boot, and optional real-time kernel support.
Multiple deployment targets: Support bare-metal, VM, and Kubernetes-native AI pipelines using ISO or RAW deployments.
Interested in learning more? Check the documentation.
Beta Was this translation helpful? Give feedback.
All reactions