Migrate access point configuration from a Classic Central instance (UI groups) to a New Central instance (sites and device groups). Provides a web UI, CLI scripts, and a Docker container.
Scope: AOS10 UI groups in Classic Central. Template groups and pre-AOS10 (Instant) groups are not supported.
Classic Central organises APs into groups. New Central uses sites (for location) and device groups (for configuration profile). This tool:
- Exports group configuration from a Classic Central tenant to disk
- Imports the APs into the equivalent New Central site and device group, using a visual mapping workflow
git clone https://github.com/<your-org>/central-group-migration.git
cd central-group-migration
docker compose up -d
open http://localhost:8000Docker downloads and installs all dependencies on first build. No other setup is required.
python3 -m venv .venv
source .venv/bin/activate # Windows: .venv\Scripts\activate
pip install -r requirements.txt
python app.py # Dev server at http://localhost:5000Both the source (Classic Central) and target (New Central) require an access token.
- In the Central UI go to Maintain → Organization → Platform Integration → API Gateway → My Apps & Tokens
- Select your application and click Generate Token
- Copy the
access_tokenvalue
Tokens are valid for 2 hours. Generate a fresh token before starting a large export or import. See DOCS.md § Authentication for the OAuth credential approach that refreshes automatically.
Cluster base URLs (common):
| Region | Base URL |
|---|---|
| US-1 | https://apigw-prod2.central.arubanetworks.com |
| US-2 | https://apigw-prod2-eu.central.arubanetworks.com |
| EU-1 | https://eu-apigw.central.arubanetworks.com |
| APAC-1 | https://apigw-apac.central.arubanetworks.com |
Your base URL is shown in Maintain → Organization → Platform Integration → API Gateway.
Open http://localhost:8000 (Docker) or http://localhost:5000 (local
dev). Three tabs:
Connects to a Classic Central source tenant and writes group configuration to disk.
- Enter the Base URL and Access Token for the source Classic Central instance
- Click Connect & Load Groups — the sidebar populates with all groups
- Select the groups to export using the checkboxes (All / None buttons are available)
- Click ↓ Export Selected
Real-time progress streams as each group is exported. Result cards show
which files were written per group (properties.json, ap_cli_config.json,
ap_inventory.json, etc.).
Assigns APs from the exported Classic Central groups to existing sites and device groups in New Central.
Click ⬡ Load Export from Disk. The sidebar populates from manifest.json
on disk. Exported groups default to no selection — check the groups you
want to import.
If labels are defined (see Data tab → Labels), a Filter by label bar appears above the group list. Click a label chip to show only groups with that label. The All and None buttons respect the active filter — they act only on visible groups. Use the Label sort option to group related groups together in the sidebar. This makes it straightforward to import one environment or location at a time.
Enter the Base URL and Access Token for the New Central target instance and click Connect to New Central. The tool fetches all existing sites and auto-matches them to exported groups by name.
The Sites mapping panel shows every selected group with a site dropdown. Groups whose name exactly matches a New Central site are auto-matched (shown with a green ✓ auto badge). Unmatched groups can be manually assigned from the dropdown.
Groups with no site mapping are skipped during import.
Click the ▶ expand button next to any group to see the APs in that group's export. Each AP row shows:
- Serial number and hostname
- Model
- IP address
- IN SITE badge (amber) — the AP is already assigned to the target site and will be automatically deselected
- IN GROUP badge (blue) — the AP is already in its expected New Central
device group (
Aruba_AP-<model>)
Use the All / None buttons or individual checkboxes to control which APs are imported. Membership data is fetched automatically from New Central once you have connected.
Below the site mapping panel, the Device Group Assignment section controls whether APs are also moved into New Central device groups after site assignment.
Enable/disable toggle — checked by default. Uncheck to skip device group assignment entirely and only perform site assignments.
How device groups are named
Device group names are derived automatically from the AP model strings
recorded in ap_inventory.json during export, using the convention:
Aruba_AP-<model>
For example, an AP-515 is placed in a group named Aruba_AP-515 and an
AP-635 goes into Aruba_AP-635. The model string comes directly from the
Classic Central monitoring API — it is the same value shown in the Central
UI device list.
Preview table
When device group assignment is enabled, a collapsible table lists every AP model present in the selected groups alongside its target device group name and the number of APs that will be moved. The table updates automatically as groups and individual APs are selected or deselected. Expand or collapse it using the summary line.
| AP Model | New Central Device Group | APs |
|---|---|---|
| AP-515 | Aruba_AP-515 | 8 |
| AP-635 | Aruba_AP-635 | 3 |
If AP detail has not yet been loaded for a group (expand the group row in the site mapping to load it), the table will prompt you to do so.
Enable Verbose logging (checkbox) to see every API call and response in the log. Click ↑ Import to New Central.
The import performs two phases:
Phase 1 — Site assignment — assigns the selected APs to their mapped
New Central site via POST /central/v2/sites/associations.
Phase 2 — Device group assignment (if enabled) — after all site assignments complete, APs are moved into model-based device groups. The tool:
- Reads AP models from
ap_inventory.jsonin each group's export directory - Checks which device groups already exist in New Central
- Creates any missing
Aruba_AP-<model>groups (New Central, AOS10, AP-only) - Checks each AP's current device group via the monitoring API
- Moves only APs not already in the correct group
A result card is shown for each site assignment and each device group assignment. Click the summary line on any card to expand the AP list:
- ✓ serial — hostname in green — AP was successfully moved
- – serial — hostname in grey — AP was already assigned (skipped)
- ✗ serial — hostname in red — AP move failed
Browse exported configuration on disk without making any API calls.
- Sidebar lists all exported groups with AP counts
- Click a group to view its detail: file inventory, group properties, AP list, and AOS10 CLI configuration (syntax-highlighted)
- Filter groups… search box filters by name (case-insensitive substring)
Labels are colored tags you can create and assign to exported groups. They let you organise groups visually and scope imports to a subset of groups without manually checking boxes one by one.
Creating labels — Type a name in the Labels panel input at the top of the Data sidebar and click + Add. Colors are assigned automatically.
Assigning labels — Each group row shows label chips below its name. Click a chip to assign it; click again to remove it. The same chips appear in the group detail panel when a group is selected.
Deleting labels — Click a chip in the Labels panel to delete it. It is removed from all groups immediately.
Label assignments are saved to exports/labels.json and persist across
restarts.
The Data tab includes a Sample Export panel for testing the import workflow without a live Classic Central source. Toggle it on, set a group name, and enter the serial numbers of real APs that exist in your New Central tenant. The sample group appears in the Import tab with a yellow TEST badge.
Disable the sample export before running a production import.
Snapshot and restore the exports/ directory. Useful before a re-export
that would overwrite existing data, or before testing a restore.
Create a backup — Click ↓ Create Backup in the sidebar. The tool
compresses the entire exports/ directory into a timestamped .tar.gz
archive (exports_YYYYMMDDTHHMMSSZ.tar.gz) stored in backups/. Requires
at least one completed export.
Restore a backup — Click ↑ Restore next to any archive. The current
exports/ directory is cleared and replaced with the archive contents. The
Data and Import tabs reflect the restored data immediately.
Delete a backup — Click ✕ Delete to permanently remove an archive. A confirmation prompt is shown.
The backup directory defaults to backups/ alongside exports/. Set the
BACKUP_DIR environment variable to change the location. In Docker, mount
it as a volume to persist backups across rebuilds:
# docker-compose.yml
volumes:
- ./exports:/app/exports
- ./backups:/app/backupsUse this tab to revert APs back to their original Classic Central groups and sites if a migration needs to be rolled back.
Assumption: Classic Central groups and sites have not been destroyed. Best practice is to leave them intact for days or weeks after confirming the migration before decommissioning.
Connect & Validate is read-only — it checks that every group and site from your export still exists in Classic Central. No APs are moved and no configuration is changed until you explicitly click Restore to Classic Central. The Classic Central credentials from the Export tab are pre-filled automatically.
After validating, the main panel shows a report for each exported group:
- ✓ Group found / ✗ Group not found — whether the Classic Central group is present
- ✓ / ⚠ Classic Central Site — whether the AP's original site is present
Warning banners highlight any missing groups or sites before you run the restore. Missing targets are skipped and reported as failures without affecting the APs that do restore successfully.
Restore to Classic Central processes each selected group in sequence:
- Moves APs back to their original Classic Central group
- Re-assigns each AP to its original Classic Central site (if recorded in the export)
A progress bar and log terminal show results in real time. Admin role is required on the token for the restore step; Connect & Validate works with a read-only token.
The CLI scripts share the same logic as the web UI and write to the same
exports/ directory.
# Export all groups from Classic Central
CENTRAL_BASE_URL=https://apigw-prod2.central.arubanetworks.com \
CENTRAL_TOKEN=<token> \
python export_groups.py
# Import groups to a target Classic Central instance
CENTRAL_BASE_URL=https://apigw-prod2.central.arubanetworks.com \
CENTRAL_TOKEN=<target-token> \
python import_groups.py# Start in background (builds on first run)
docker compose up -d
# Rebuild after code changes
docker compose up -d --build
# View live logs
docker compose logs -f
# Stop (exported data in ./exports/ is preserved)
docker compose downData persists across restarts and rebuilds.
By default, exports and backups are stored in exports/ and backups/
alongside the project directory — this works on all platforms with no
configuration required.
| Platform | What to do |
|---|---|
| macOS (Docker Desktop) | Nothing — Docker Desktop handles bind-mount permissions automatically. |
| Windows (Docker Desktop) | Nothing — same as macOS. |
| Linux workstation | Works out of the box with the default ./exports path. |
| Linux remote server | Set EXPORT_HOME and PUID/PGID in .env (see below). |
When the project lives in a system directory (e.g. /opt/) it can be more
convenient to store exported data in your home directory. Set these in .env:
# .env — Linux remote server example
EXPORT_HOME=/home/dan # host path used as the volume root
PUID=1000 # match `id` output on the host
PGID=1000Exports land at $EXPORT_HOME/central-exports/ and backups at
$EXPORT_HOME/central-backups/ on the host. PUID/PGID remap the
container user to match the owner of that directory, preventing Permission
denied errors.
sudo caveat — Running
docker composeundersudochanges$HOMEto/root. If you use sudo, either runsudo -E docker compose up -d(which preserves your environment) or setEXPORT_HOMEexplicitly in.envso the path does not depend on$HOME.
Set HOST_PORT in .env:
# .env
HOST_PORT=9090Then restart: docker compose up -d. No edits to docker-compose.yml needed.
Copy .env.example to .env as a starting point if the file doesn't exist yet.
central-group-migration/
├── app.py # Flask backend (routes, SSE, orchestration)
├── exporters.py # Per-data-type export/import registry
├── new_central_importer.py # New Central site and device group import logic
├── export_groups.py # CLI export script
├── import_groups.py # CLI import script
├── templates/
│ └── index.html # Web UI — Export / Import / Data tabs
├── exports/ # Runtime data (bind-mounted, git-ignored)
├── Dockerfile
├── docker-compose.yml
├── requirements.txt
├── README.md # This file
└── DOCS.md # Full technical documentation
| Package | Version | Purpose |
|---|---|---|
pycentral |
≥ 2.0a17 | Aruba Central Python SDK — all API calls |
flask |
3.1.0 | Web framework |
flask-cors |
5.0.0 | CORS headers |
gunicorn |
23.0.0 | Production WSGI server (single worker, 8 threads) |
See DOCS.md for:
- Architecture and request flow
- Docker build verification
- Authentication (access tokens and OAuth)
- Complete web UI guide
- CLI scripts guide
- Export file format and JSON schemas
- New Central API surface reference
- HTTP API reference with SSE event payloads
- Troubleshooting
- Extending the tool with new exporters
MIT