Replace the machine. Keep the project.
Delete a machine without losing its project files. Follow a managed volume through a controller restart, a replacement machine, read-only access and explicit cleanup.
Read the guideRunning code
beyond the BEAM.
Practical notes on Python, JavaScript, and the execution lifecycle in SmolBox.
Engineering notes
Subscribe via RSS
Exports, checkpoints or branches: what should you reuse?
Choose what your next machine inherits. Compare disk artifacts, saved memory and live copies, with real Linux measurements and the cleanup each choice leaves behind.
Read the guide
Your browser disconnected. Is the command still running?
Follow a terminal session through a lost connection. Learn when you can reconnect, what an exit proves, and why an uncertain command should not be replayed.
Read the guide
Your VM is running. Is your application ready?
Start an HTTP service, break its startup command, and check what a running VM actually tells you. A practical SmolBox guide to application readiness.
Read the guide
Build a persistent workspace with Phoenix and SmolBox 0.2.0
Edit a page, run successive commands, and restart Phoenix without losing the project. A runnable workspace teaches machine, command, and browser lifetimes.
Read the guide
Prepare once, run in a fresh VM with SmolBox
Restore prepared guest state for separate disposable executions. A station dataset shows what checkpoints preserve, what they cost, and when a prepared image is enough.
Read the guide
How much slower is nested KVM? Measuring our SmolBox lab
A reader asked about nested virtualization overhead. Six paired tests showed why startup, guest execution and cleanup need separate clocks.
Read the guide
Controlled network access for disposable VMs in Elixir
Fetch a public dataset from a disposable Python VM, collect the report in Elixir, and keep its approved network policy with the execution. A working SmolBox 0.1.3 example.
Read the guide
From a full disk to an upstream fix
A cleanup failure in our Elixir integration, a fix from smolvm's maintainer, and the tests that closed the loop. A small story of contributing to open source.
Read the guide
When the sandbox fails: testing SmolBox with nested KVM
What disk exhaustion, worker OOM and a frozen VM taught us about resource limits, execution records and recovery in an Elixir application.
Read the guide
Introducing SmolBox: managing smolvm executions from Elixir
Why we chose smolvm and how SmolBox manages execution identity, outcomes, and cleanup from Elixir, with a runnable Python example.
Read the guide