We’re currently packaging the Polkadot binary as a Snap to make it easier for operators to install and maintain.
When testing validator mode inside the Snap, we’ve hit a blocker: Polkadot attempts to enable extra process isolation via Linux Landlock, but the landlock_restrict_self syscall fails in the Snap environment. This happens because Snaps are already confined by AppArmor and seccomp, which makes Landlock unavailable.
Our concern is that this forces users to run with the flag:
--insecure-validator-i-know-what-i-do
which weakens security guarantees.
Questions for the Snap team:
- Is there any way to expose Landlock capabilities within a Snap, or have the Snap confinement recognized as an equivalent?
- If direct Landlock support isn’t possible, what’s the recommended approach to package Polkadot securely as a Snap without relying on the insecure validator flag?
- Would it make sense to handle this at the Snap level (e.g., through an interface, confinement tweak, or guidance on security model alignment)?
Our goal is to ship a secure, confined Snap for Polkadot validators without compromising on safety. Any advice or suggested next steps from the Snap side would be very helpful.
Thanks!