v2.0.0

Notice

ZTF 2.0.0 is a ground-up rewrite using the Nutanix v4 API Python SDK. The legacy
v2 / v3 / Foundation Central / NCM-Calm / NDB workflows from the 1.x line
are not yet ported (see Migrating from ZTF 1.x
in the project README); 2.x establishes the new declarative,
multi-domain, plan-and-apply foundation.

What's New

Framework Rewrite

• New declarative configuration model: input.yml (resources & data
  sources) + global.yml (SDK / connection settings).
• New ztf Click CLI replacing the legacy python main.py --workflow
  invocation. Commands: init, plan, apply, refresh, destroy,
  import, examples, repl.
• Multi-domain orchestration: a single run can target many Prism
  Centrals in parallel, each with its own data sources, resources,
  and credentials.
• Dependency resolution: automatic topological ordering via
  interpolation analysis ({resource_name.extId},
  {domain.resource.field}) plus explicit depends_on.
• State management: YAML state file with advisory file locking,
  automatic backups, and incremental writes.
• Auto-refresh: plan, apply, and destroy reconcile state with
  live infrastructure first; --no-refresh and --strict available
  for CI flows.
• Confirmation prompt before apply / destroy; --auto-approve
  for trusted CI/CD.
• Post-apply summary with created / updated / deleted / failed counts.

v4 SDK Coverage

• 16 Nutanix v4 namespaces wired through ntnx-*-py-client packages:
  clustermgmt, networking, prism, vmm, iam, volumes,
  microseg, dataprotection, monitoring, lifecycle, aiops,
  security, licensing, objects, opsmgmt, datapolicies,
  multidomain.
• Schema negotiation: per-PC-version request body filtering driven by
  multi_namespace_compatibility_map.json so the same input.yml
  works across PC minors that ship different namespace versions.
• ETag / If-Match round-trip handled automatically via SDK reserved
  fields and a curated 268-entity etag_required map.
• NTNX-Request-Id auto-injection on every mutation (SDK-level).

Plan / Apply Lifecycle

• Field-level diff output in plan showing exactly which fields will
  change on update.
• Plan-file workflow: ztf plan --out plan.json writes a canonical
  plan; ztf apply --plan plan.json [--force] replays it with hash
  verification to guarantee plan/apply consistency.
• Selective destroy via --target (repeatable).
• Cross-domain reverse-wave destroy: dependent domains tear down
  before the domains they reference.
• --strict makes refresh errors a hard failure (CI-friendly).

Configuration Surface

• Variables and var-files with multi-source override precedence.
• Auto-loaded variable files: ztfvars.yml and *.auto.ztfvars.yml.
• Dynamic resources: for_each expansion plus the new count: N
  sugar (desugars to for_each: {str(i): i} with each.index).
• Lifecycle rules: prevent_destroy, ignore_changes,
  create_before_destroy (beta).
• Data sources: read-only lookups of existing infrastructure with
  filter support; resolved at apply time.
• Outputs section with YAML / JSON file export via --output-file.
• User-defined {fn.*} functions (functions.py) for computed
  values, templates, and base64 encoding.
• Interpolation tokens: {var.name}, {each.key}, {each.value},
  {each.index}, {fn.name(args)}, {data.entity.name.field},
  {resource_name.extId}, {domain.resource.field}.

Import & Examples

• ztf import <resource> <domain> brings existing infrastructure under
  state with a suggested config snippet printed for the user.
• ztf import --file <manifest> for bulk import.
• ztf examples auto-generates per-entity YAML examples and Markdown
  reference docs (field descriptions, enum values, version
  availability) under examples/ plus a starter functions.py.
• ztf repl opens a read-only Python REPL with data sources loaded
  for ad-hoc value discovery.

Operational Improvements (Phase A–E hardening)

Drop-in fixes from the ztf-2.0-rc branch — no breaking changes:

• Sub-entity updates correctly resolve referenced parent ext-ids and
  dispatch the right SDK method, fixing update drift on nested
  resources.
• run_results/ resolves relative to the invoking workspace instead
  of the package install path.
• Deepcopy guard prevents mutation of cached entity payloads across
  repeated plan / apply invocations.
• Entity cache keys use tuple composition instead of string
  concatenation, eliminating cross-entity cache-key collisions.
• Deep structural comparison rewritten from quadratic
  match-and-delete to a linear multiset freeze; significantly faster
  plan on large states.
• Internal tech-debt cleanup in provider.py and
  entity_handler.py.

Security

• setuptools.packages.find restricts wheel contents to the ztf*
  namespace, preventing tests/, scripts/, examples/, and
  workflow/ from shipping as importable top-level packages on
  install.
• License switched from MIT (1.x) to Apache-2.0; see LICENSE.txt
  and NOTICE.

Tooling

• Python 3.10, 3.11, 3.12, and 3.13 supported.
• Packaged via pyproject.toml (legacy setup.py removed); managed
  with uv and published to PyPI as nutanix-ztf.
• Multi-stage Dockerfile shipping a runtime image and a
  wheels-export target for dark-site offline installs.
• New CI surface: ci.yml, publish.yml, functional-tests.yml,
  black-duck.yml. Linting & formatting via ruff, type checking
  via mypy, security scanning via bandit.

Known Limitations

• Nutanix Files APIs not supported due to an upstream package
  dependency conflict; tracking with Nutanix.
• update, delete, create_before_destroy, operations, and
  destroy are functional but beta — always run ztf plan
  before ztf apply and review the diff.
• Legacy 1.x feature surfaces (Foundation Central imaging, Prism
  Element v2 operations, Pod / pod-block conceptual workflows, NCM /
  Calm DSL workloads, NDB, NKE / Karbon, legacy script-based Objects
  flow) are not yet ported. Use the
  1.x branch
  (last release v1.5.2) until parity lands. See
  Migrating from ZTF 1.x.

Migration

Anyone tracking nutanixdev/zerotouch-framework@main on a pinned
SHA will see a large delete-and-rewrite as 2.0.0 ships:
framework/, calm-dsl-bps/, calm-whl/, requirements/,
dev-setup-README.md, setup.py, main.py, and
config/example-configs/ are removed; the new tree is rooted at the
ztf/ Python package plus pyproject.toml. Pin to v1.5.2 or the
1.x branch if you need the legacy surfaces. Forward parity work is
ongoing.

v1.5.2

What's New

Bug Fixes

• Fixed Issue related to Hypervisor Type Field
• Fixed Issue with CMSP Timeout
• Fixed Issue with Enable Network Controller
• Fixed Issue with IPMI Gateway not being Optional
• Fixed Issue with PC 7.x version because of API Version Switching

1.5.1

What's New

Bug Fixes

• Fixed issue with 'missing' commits in the release pipeline
• Added 'IAM' v4 client in requirements.txt

v1.5.0

What's New

Framework Enhancements

• New site-deploy workflow for deploying sites leveraging Foundation Central
• New cluster-create and imaging-only workflows which also leverage Foundation Central

New PC Operations

• Add IAM entities in PC using v4 APIs, which include create ops for Roles, UserGroups, Users, IAMKeys, AuthorizationPolicy, DirectoryServices
• Enable MarketPlace
• Enable Foundation Central
• Generate Foundation Central API keys

New PE operations

With the v2 API endpoint for virtual switch being deprecated we've added the v4 API endpoint as part of the create_pe_subnets_script

Bug Fixes

• Fixed several bugs in create_pc_vms script
• Fixed a bug related to ipmi_netmask field in foundation_script
• Fixed a bug in create_objectstores script

v1.4.1

What's New

Framework Enhancements

• Integrated Unit Test Framework
• Introduced v4 API SDK for PC
• Added backward compatibility for v3 APIs for older PC versions by implementing a mechanism to choose the API version based on the PC build version

Initial NDB Functionality

• Deploy & Configure NDB Management
  • Deploy NDB VMs
  • Change NDB VM password
  • Register NDB Clusters
  • Enable NDB Multi-Cluster
  • Create Compute Profiles
  • Create Network Profiles
  • Enable NDB HA
  • Enable / Disable Pulse

New Operations Functionalities

Imaging Workflow Update

• Added support to specify VLAN ID in the imaging workflow

PE Operations

• Upload Image to PE
• Create VM in PE
• Power VM On/Off in PE

PC Deployment Operations

• Deploy multiple PC instances to a single Nutanix Cluster
• Select which PC to register with PE
• Enable and configure CMSP during PC deployment

PC Configuration Operations

• Enabled/Disabled Network Controller in PC
• Create VMs and manage power transitions
• New v4 API CRUD Operations for:
  • Categories
  • Address Groups
  • Service Groups
  • Security Policies (Flow Network Security Next-Gen)
  • VPCs

Bug fixes

• Fixed an issue where hypervisor_hostname was being ignored in the API payload
• Fixed an issue where the IPMI gateway was being passed instead of the IPMI subnet

v1.3.1

v1.3.1

CVM Foundation version enhancement

• An additional script has been included to modify the CVM Foundation version (either upgrade or downgrade)

Create Identity Provider (IDP) in IAM

• We now facilitate the setup of any IDPs compliant with SAML 2.0 in Prism Central

Hypervisor hostname Configuration in Imaging

• "hypervisor_hostname" can now be defined in pod-deploy.yml
  to configure the Hypervisor hostname during Imaging

Objects

• It is now possible to define "storage_network" and "public_network" separately when creating an Objectstore

Cyberark

• The Cyberark integration has been enhanced to support the Cyberark AIM API

Cluster Configuration workflow

• HA reservation and Rebuild Capacity Reservation are now supported in the Cluster Configuration workflow

v1.3.0.2

v1.3.0.2

What's New

Improvement

• The framework is now compatible with Python 3.9 and higher, as opposed to its previous compatibility with Python 3.10

v1.3.0.1

v1.3.0.1

What's New

Bug

• Fixed invalid schema bug for "deploy-pc" workflow

v1.3.0

v1.3.0

What's New

Framework Enhancements

• Integration with CyberArk Vault for secure storage. Additional vault integrations will be included in future updates.
• Integration with Infoblox IPAM for efficient IP address management. Support for other IPAM solutions will be added in upcoming iterations.
• Centralized storage of credential and IPAM configurations in global.yml file.
• Introduction of new workflows:
  • config-cluster: Configures the cluster settings.
  • deploy-pc: Deploys the PC (Prism Central) entity.
  • config-pc: Configures the PC entities.
• ZTF now supports delete operations on both PC and PE entities.
• Improved PC deployment: If a file failed to download previously, the download process will now resume instead of starting from scratch.

Imaging and Cluster Creation Features

• Integration with IPAM for Imaging and Cluster deployment.
• Ability to specify CVM, Host, and IPMI IPs for each node individually.

v1.2.1.2

v1.2.1.2

What's New

Bug

• Fixed Module Not Found Error for PCDeploy
• Updated example config