Getting started with GLPI inventory: first steps for IT managers

Getting started with GLPI inventory: first steps for IT managers

GLPI installs in half an hour. Filling it with data so it acts as a real inventory takes a day. Actually keeping it accurate over time is a process question. This walks through the first concrete steps: how to set up the data so you can pour your existing hardware and software into GLPI without redoing everything a month later.

Structure first, then data

The most common mistake — pouring assets into GLPI right after install — catches up with you at the first reorg. Before adding the first computer, configure these four foundational dimensions:

  1. Entities (Administration → Entities) — which organizational unit owns the assets. At least one entity below Root, even if it's just "Main Company". Without entities you cannot later separate data between teams or clients (more in entity-based RBAC).
  2. Locations (Setup → Dropdowns → Locations) — physical place. Hierarchical: "Bratislava" → "Bratislava HQ" → "Bratislava HQ 3rd floor". An asset's location later attaches to tickets and simplifies field-service decisions.
  3. Groups (Administration → Groups) — who's responsible for which assets. "Servers — Linux", "Servers — Windows", "Printers", "Laptop fleet".
  4. Users (Administration → Users) — who assets are assigned to. Import from LDAP/AD before loading assets; assets attached to nonexistent users mean manual repair work later.

Asset types and dropdowns

GLPI ships with standard asset types: Computer, Monitor, Network device, Printer, Phone, Software, Cartridge, Consumable. They cover 95 % of deployments. If you need to track anything that doesn't fit those categories (cloud-service licenses, IoT sensors), install the Generic Object plugin and define your own type.

For each type GLPI expects dropdowns: manufacturers, models, OS versions, processor types, monitor categories, etc. A fresh install starts empty — they fill in either manually or automatically via GLPI Agent once you start network discovery.

One practical rule: don't use free-text fields where a dropdown exists. Free text creates variants ("Dell Latitude E5570", "Dell Latitude e5570", "DELL Latitude E5570") and reporting never gives a clean count.

Manual entry vs CSV import

Three options for loading an existing inventory:

  • Manual entry — works for tens of assets, painful for hundreds. Useful for a sanity test of your structure.
  • CSV import via the Data Injection plugin — the recommended path for hundreds to thousands of assets. You prepare a CSV with a header matching GLPI fields (name,serial,otherserial,type,location,users_id...) and the importer creates records, including auto-creating missing dropdown values on demand.
  • Automatic discovery via GLPI Agent — the agent on the target device pushes inventory to GLPI. Best path for continuous freshness, but deploying agents is a project of its own.

For most new implementations the best combination is: a one-time CSV import from the spreadsheet you already keep, plus a parallel GLPI Agent rollout that maintains the data from then on.

CSV import: required header fields

Data Injection maps CSV columns to GLPI fields by header name. A minimal usable header for computers:

name,serial,manufacturers_id,computermodels_id,locations_id,users_id,states_id,entities_id
PC-001,SN12345,Dell,Latitude E7470,BA-3rd,jane.doe,In production,Main Company
PC-002,SN12346,Lenovo,ThinkPad T14,BA-3rd,john.smith,In production,Main Company

states_id is the asset status — see asset states and transitions in GLPI for lifecycle configuration. entities_id determines who can see the asset — get the entity wrong and the asset effectively disappears from your team's view.

First-week checklist

  • Day 1: configure entities, locations, groups. Import users from LDAP/AD.
  • Day 2: define which asset types you'll track and seed the key dropdowns (manufacturers, models for the most common devices).
  • Day 3: prepare a CSV template, test the import on 5–10 records.
  • Day 4: bulk CSV import.
  • Day 5: review imported data, fix missing assignments, and define three baseline saved searches: Assets without an owner, Assets out of warranty, Assets in Repair status.

After a week you have a working inventory. The next step is configuring how assets move through their lifecycle — from intake to retirement.

Need help with this topic?

Get in touch