C
CYSKA
Consulting
EN FR
Navigation
OpenStack

Modernize OpenStack without putting production at risk

We assess your release gap, storage and workloads, then choose the safest path: parallel platform, controlled export/import or sequential in-place upgrade. Every path includes testing, rollback criteria and production stabilization.

Migration path
Legacy → Supported
  1. 1
    Audit and inventory
    Releases, storage, networks, workloads and their acceptable outage windows.
  2. 2
    Parallel target platform
    Built with Kolla-Ansible next to the legacy cloud, sharing the Ceph cluster when possible.
  3. 3
    Pilot group
    Validates the procedure, acceptance criteria and rollback before critical workloads move.
  4. 4
    Controlled waves and cutover
    Volumes adopted by the new Cinder without copying data; legacy cloud kept as rollback until acceptance.
See the technical detail, command by command →
Approach

Three migration paths for legacy OpenStack environments

Option A: Shared Ceph cluster

The Ceph pool is connected to both old and new OpenStack deployments. Volumes are unmanaged from legacy Cinder and reimported into the new Cinder service without network copy.

Option B: Snapshot, export and import

For isolated storage backends, we create snapshots, export images to Glance/S3, move them, and recreate the environment in the target OpenStack cluster.

Option C: In-place upgrade (kolla-ansible upgrade)

The existing cluster is upgraded in place through a path supported by Kolla-Ansible and every deployed service, without deploying a parallel environment. Faster and cheaper, but it carries real risk for the hosted workloads.

Options A and B are documented command by command in our technical guide: Migrate Cinder volumes between two OpenStack clouds →

In-place

In-place upgrade with kolla-ansible

kolla-ansible upgrade rolls newer container images out on the existing multi-AZ cluster without building a parallel target. Use adjacent releases, or a skip between two consecutive SLURP releases only when Kolla-Ansible and every deployed service support it. A legacy source can therefore require several qualified maintenance windows.

Technical guide: release-by-release upgrade procedure →

We do not recommend this path by default

An in-place Kolla-Ansible upgrade changes the live control plane and compute services. A failed step can extend the maintenance window or disrupt APIs and workloads. We only retain this path after validating backups, rollback options, application resilience, compatibility between each release and explicit risk acceptance.

01

Audit and inventory

We map services, storage, network, and OpenStack versions before proposing the migration path.

02

Parallel deployment

We build the target platform alongside the legacy environment to minimize business disruption.

03

Cutover and control

We validate workloads, ensure HA, and stabilize production before final cutover.

What your team receives

A version and dependency inventory, tested migration runbooks, pilot report, per-wave cutover and rollback plans, acceptance evidence, updated architecture documentation and knowledge transfer.

Assess my OpenStack migration