Quick Start¶

Get your network visible to OLAV in 5 minutes: install, configure devices, collect data, ask your first question.

Feature Claims

ID	Claim	Status
C-NE-01	`/netops_init --dry-run` validates environment + Nornir inventory	✅ v0.17.0
C-NE-02	`olav list` shows the 6 platform agents (admin, audit, core, devops, netops, services) after `olav agent install`	✅ v0.18.0
C-NE-03	`/netops_init` populates `netops.parsed_outputs` with platform-aware commands	✅ v0.17.0
C-NE-04	`/netops_init` generates topology + populates `netops.devices`	✅ v0.17.0

Legend: ⚠️ (E2E-gated) = test exists but requires env-var flag, skipped in CI; ⚠️ (manual) = demo-verified only, no automated CI test

Step 1: Install OLAV + olav-netops¶

pip install olav olav-netops     # OLAV core + the network domain package
olav agent install olav-netops   # deploy the netops workspace and register the agent

pip install pulls both packages from PyPI. olav agent install olav-netops then resolves the installed package, deploys its workspace (the netops agent plus network-aware audit), and refreshes the agent routing table. (For development, point olav agent install at a local path or git URL instead: olav agent install /path/to/olav-netops/.)

First run bootstraps core

On a brand-new project, run olav once first — it scaffolds the workspace, databases, and core agent, and prompts for your LLM API key. See Installation for the full core setup.

Verify:

olav list

Expected output — 6 top-level agents:

  • admin     — Platform self-management: health/log diagnostics, cron, skill install, workspace scaffolding
  • audit     — Compliance checks, drift detection, network health report generation
  • core      — Cross-domain: SQL, semantic recall, Python sandbox, web search, reports, cron
  • devops    — NetBox DCIM/IPAM, InfluxDB metrics, automation script library, HITL execution
  • netops    — Network troubleshooting: topology analysis, routing simulation, lab validation, probing
  • services  — Service integration: register APIs, deploy/stop containers, authenticated HTTP calls

Step 2: Configure Your Devices¶

Device inventory¶

cp .olav/workspace/netops/collector/config/nornir/hosts.yaml.example \
   .olav/workspace/netops/collector/config/nornir/hosts.yaml
nano .olav/workspace/netops/collector/config/nornir/hosts.yaml

Example hosts.yaml:

R1:
  hostname: 192.168.100.101
  platform: juniper_junos
  groups:
    - border
    - juniper_junos

R2:
  hostname: 192.168.100.102
  platform: cisco_ios
  groups:
    - core
    - cisco_ios

SSH credentials¶

cp .olav/workspace/netops/collector/config/nornir/defaults.yaml.example \
   .olav/workspace/netops/collector/config/nornir/defaults.yaml
chmod 600 .olav/workspace/netops/collector/config/nornir/defaults.yaml
nano .olav/workspace/netops/collector/config/nornir/defaults.yaml

username: your_ssh_username
password: your_ssh_password
port: 22

Warning

Nornir requires plain text passwords. defaults.yaml is gitignored — never commit it.

Step 3: Run Initialization¶

olav --agent netops "/netops_init"

This runs a 4-stage pipeline:

Stage	What happens
1. Environment Check	Validates Nornir config, SSH packages, inventory (group references verified)
2. SSH Collection	Platform-aware commands: Junos devices get `show lldp neighbors`, `show interfaces terse`, etc.; IOS devices get `show cdp neighbors detail`, `show ip interface brief`, etc.
3. Ingest + ETL	IngestManager writes to DuckDB → Topology ETL extracts CDP/LLDP links → Device ETL extracts hostname, IP, vendor, model, OS version
4. Summary	Reports devices collected, commands, success/failure count, topology links

Dry run first

Validates environment without SSH connections:

olav --agent netops "/netops_init --dry-run"

Checks: Nornir installed, hosts.yaml exists, group names match groups.yaml, inventory loads without error.

Platform-aware collection

The pipeline automatically selects commands based on device platform:

Cisco IOS: 14 commands (show ip interface brief, show cdp neighbors detail, show running-config, ...)
Juniper JunOS: 11 commands (show interfaces terse, show lldp neighbors, show configuration, ...)
Universal: show version, show interfaces (all platforms)

Step 4: Ask Your First Question¶

olav "How many devices are in my network?"

olav "Show me the topology"

olav "Are all BGP neighbors established?"

OLAV queries the collected data in DuckDB and returns structured answers. No device access is needed for queries — everything comes from the snapshot database.

What's Next?¶

Basic Concepts — Understand how snapshots, templates, and agents work
Core Agent — Daily queries and CLI execution
NetOps Agent — Deep analysis, simulation, and drift detection