How to Install & Set Up HashiCorp Vault
Install HashiCorp Vault from the official APT repository, write your first secret, then see how it initializes/unseals/authorizes for real once you're past the dev server.
Prerequisites
- ✓A Linux host
What Vault is actually managing
Vault sits between callers and secrets and mediates every request through two layers: auth methods, which establish who's calling (tokens, LDAP, a Kubernetes service account, AWS IAM), and policies, HCL rules that say what paths that identity can read or write. The secrets themselves are organized behind secrets engines — a static key-value store for the simple case, but also database, PKI, and cloud engines that mint short-lived, automatically-expiring credentials on demand instead of handing out something permanent.
Install on Ubuntu/Debian (HashiCorp APT repo)
Install on RHEL, CentOS Stream, Fedora, or Amazon Linux (HashiCorp YUM repo)
Alternative — run Vault in Docker
Start a dev server
Vault is licensed under the BUSL. For a fully open source, OSI-approved alternative with the same CLI, see the OpenBao guide.
Write and read a secret
Beyond -dev: init and unseal
A dev server starts already unsealed and pre-authenticated, which is exactly what makes it unsuitable for anything real. A production Vault generates its master encryption key once, during vault operator init, and immediately splits it into key shares (5 by default, needing any 3 to reconstruct) using Shamir's Secret Sharing — no single operator ever holds the whole key. Every restart leaves Vault sealed until enough shares are supplied again.
Vault cannot decrypt its data if you lose enough key shares to drop below the threshold — there's no backdoor. Store the shares separately from the server and from each other.
Scope access with a policy
Documentation & Knowledge Base
Download HashiCorp Vault— all versions & official links→Need HashiCorp Vault deployed and managed for you?
We're OpenStack Dev — we design, deploy, and operate production DevOps infrastructure.