Files
vault/enos/modules/start_vault/main.tf
Ryan Cragun a087f7b267 [QT-627] enos: add pkcs11 seal testing with softhsm (#24349)
Add support for testing `+ent.hsm` and `+ent.hsm.fips1402` Vault editions
with `pkcs11` seal types utilizing a shared `softhsm` token. Softhsm2 is
a software HSM that will load seal keys from a local disk via pkcs11.
The pkcs11 seal implementation is fairly complex as we have to create a
one or more shared tokens with various keys and distribute them to all
nodes in the cluster before starting Vault. We also have to ensure that
each sets labels are unique.

We also make a few quality of life updates by utilizing globals for
variants that don't often change and update base versions for various
scenarios.

* Add `seal_pkcs11` module for creating a `pkcs11` seal key using
  `softhsm2` as our backing implementation.
* Require the latest enos provider to gain access to the `enos_user`
  resource to ensure correct ownership and permissions of the
  `softhsm2` data directory and files.
* Add `pkcs11` seal to all scenarios that support configuring a seal
  type.
* Extract system package installation out of the `vault_cluster` module
  and into its own `install_package` module that we can reuse.
* Fix a bug when using the local builder variant that mangled the path.
  This likely slipped in during the migration to auto-version bumping.
* Fix an issue where restarting Vault nodes with a socket seal would
  fail because a seal socket sync wasn't available on all nodes. Now we
  start the socket listener on all nodes to ensure any node can become
  primary and "audit" to the socket listner.
* Remove unused attributes from some verify modules.
* Go back to using cheaper AWS regions.
* Use globals for variants.
* Update initial vault version for `upgrade` and `autopilot` scenarios.
* Update the consul versions for all scenarios that support a consul
  storage backend.

Signed-off-by: Ryan Cragun <me@ryan.ec>
2023-12-08 14:00:45 -07:00

242 lines
8.3 KiB
HCL

# Copyright (c) HashiCorp, Inc.
# SPDX-License-Identifier: BUSL-1.1
terraform {
required_providers {
# We need to specify the provider source in each module until we publish it
# to the public registry
enos = {
source = "app.terraform.io/hashicorp-qti/enos"
version = ">= 0.4.8"
}
}
}
locals {
bin_path = "${var.install_dir}/vault"
environment = local.seal_secondary == null ? var.environment : merge(
var.environment,
{ VAULT_ENABLE_SEAL_HA_BETA : tobool(var.seal_ha_beta) },
)
// In order to get Terraform to plan we have to use collections with keys that are known at plan
// time. Here we're creating locals that keep track of index values that point to our target hosts.
followers = toset(slice(local.instances, 1, length(local.instances)))
instances = [for idx in range(length(var.target_hosts)) : tostring(idx)]
leader = toset(slice(local.instances, 0, 1))
// Handle cases where we might have to distribute HSM tokens for the pkcs11 seal before starting
// vault.
token_base64 = try(lookup(var.seal_attributes, "token_base64", ""), "")
token_base64_secondary = try(lookup(var.seal_attributes_secondary, "token_base64", ""), "")
// This module currently supports up to two defined seals. Most of our locals logic here is for
// creating the correct seal configuration.
seals = {
primary = local.seal_primary
secondary = local.seal_secondary
}
seals_primary = {
awskms = {
type = "awskms"
attributes = merge(
{
name = var.seal_alias
priority = var.seal_priority
}, var.seal_attributes
)
}
pkcs11 = {
type = "pkcs11"
attributes = merge(
{
name = var.seal_alias
priority = var.seal_priority
},
// Strip out attributes that aren't supposed to be in seal stanza like our base64 encoded
// softhsm blob and the token directory. We'll also inject the shared object library
// location that we detect on the target machines. This allows use to create the token and
// keys on a machines that have different shared object locations.
merge(
try({ for key, val in var.seal_attributes : key => val if key != "token_base64" && key != "token_dir" }, {}),
try({ lib = module.maybe_configure_hsm.lib }, {})
),
)
}
shamir = {
type = "shamir"
attributes = null
}
}
seal_primary = local.seals_primary[var.seal_type]
seals_secondary = {
awskms = {
type = "awskms"
attributes = merge(
{
name = var.seal_alias_secondary
priority = var.seal_priority_secondary
}, var.seal_attributes_secondary
)
}
pkcs11 = {
type = "pkcs11"
attributes = merge(
{
name = var.seal_alias_secondary
priority = var.seal_priority_secondary
},
merge(
try({ for key, val in var.seal_attributes_secondary : key => val if key != "token_base64" && key != "token_dir" }, {}),
try({ lib = module.maybe_configure_hsm_secondary.lib }, {})
),
)
}
none = {
type = "none"
attributes = null
}
}
seal_secondary = local.seals_secondary[var.seal_type_secondary]
storage_config = [for idx, host in var.target_hosts : (var.storage_backend == "raft" ?
merge(
{
node_id = "${var.storage_node_prefix}_${idx}"
},
var.storage_backend_attrs
) :
{
address = "127.0.0.1:8500"
path = "vault"
})
]
}
# You might be wondering why our start_vault module, which supports shamir, awskms, and pkcs11 seal
# types, contains sub-modules that are only used for HSM. Well, each of those seal devices has
# different requirements and as such we have some seal specific requirements before starting Vault.
#
# A Shamir seal key cannot exist until Vault has already started, so this modules responsibility for
# shamir seals is ensuring that the seal type is passed to the enos_vault_start resource. That's it.
#
# Auto-unseal with a KMS requires that we configure the enos_vault_start resource with the correct
# seal type and the attributes necessary to know which KMS key to use. Vault should automatically
# unseal if we've given it the correct configuration. As long as Vault is able to access the key
# in the KMS it should be able to start. That's normally done via roles associated to the target
# machines, which is outside the scope of this module.
#
# Auto-unseal with an HSM and PKCS#11 is more complicated because a shared object library, which is
# how we interface with the HSM, must be present on each node in order to start Vault. In the real
# world this means an actual HSM in the same rack or data center as every node in the Vault cluster,
# but in our case we're creating ephemeral infrastructure for these test scenarios and don't have a
# real HSM available. We could use CloudHSM or the like, but at the time of writing CloudHSM
# provisioning takes anywhere from 30 to 60 minutes and costs upwards of $2 dollars an hour. That's
# far too long and expensive for scenarios we'll run fairly frequently. Instead, we test using a
# software HSM. Using a software HSM solves the cost and speed problems but creates new set of
# problems. We need to ensure every node in the cluster has access to the same "HSM" and with
# softhsm that means the same software, configuration, tokens and keys. Our `seal_pkcs11` module
# takes care of creating the token and keys, but that's the end of the road for that module. It's
# our job to ensure that when we're starting Vault with a software HSM that we'll ensure the correct
# software, configuration and data are available on the nodes. That's where the following two
# modules come in. They handle installing the required software, configuring it, and distributing
# the key data that was passed in via seal attributes.
module "maybe_configure_hsm" {
source = "../softhsm_distribute_vault_keys"
hosts = var.target_hosts
token_base64 = local.token_base64
}
module "maybe_configure_hsm_secondary" {
source = "../softhsm_distribute_vault_keys"
depends_on = [module.maybe_configure_hsm]
hosts = var.target_hosts
token_base64 = local.token_base64_secondary
}
resource "enos_vault_start" "leader" {
for_each = local.leader
depends_on = [
module.maybe_configure_hsm_secondary,
]
bin_path = local.bin_path
config_dir = var.config_dir
environment = local.environment
config = {
api_addr = "http://${var.target_hosts[each.value].private_ip}:8200"
cluster_addr = "http://${var.target_hosts[each.value].private_ip}:8201"
cluster_name = var.cluster_name
listener = {
type = "tcp"
attributes = {
address = "0.0.0.0:8200"
tls_disable = "true"
}
}
log_level = var.log_level
storage = {
type = var.storage_backend
attributes = ({ for key, value in local.storage_config[each.key] : key => value })
}
seals = local.seals
ui = true
}
license = var.license
manage_service = var.manage_service
username = var.service_username
unit_name = "vault"
transport = {
ssh = {
host = var.target_hosts[each.value].public_ip
}
}
}
resource "enos_vault_start" "followers" {
depends_on = [
enos_vault_start.leader,
]
for_each = local.followers
bin_path = local.bin_path
config_dir = var.config_dir
environment = local.environment
config = {
api_addr = "http://${var.target_hosts[each.value].private_ip}:8200"
cluster_addr = "http://${var.target_hosts[each.value].private_ip}:8201"
cluster_name = var.cluster_name
listener = {
type = "tcp"
attributes = {
address = "0.0.0.0:8200"
tls_disable = "true"
}
}
log_level = var.log_level
storage = {
type = var.storage_backend
attributes = { for key, value in local.storage_config[each.key] : key => value }
}
seals = local.seals
ui = true
}
license = var.license
manage_service = var.manage_service
username = var.service_username
unit_name = "vault"
transport = {
ssh = {
host = var.target_hosts[each.value].public_ip
}
}
}
output "token_base64" {
value = local.token_base64
}
output "token_base64_secondary" {
value = local.token_base64_secondary
}