Labtris docs

Organising labs

Folders, clone, lock, config sets, import and export.

One lab is easy. Forty is the problem this page is about.

Folders

Every lab has a folder, which is just a path string — ccna/week-3, customer/acme, scratch. The lab list groups by it, and the breadcrumb in the header shows where you are.

Move a lab by editing its folder in the inspector. Renaming a folder moves everything in it:

curl -X POST http://<host>:8081/api/v1/folders/rename \
  -H 'content-type: application/json' \
  -d '{"from": "ccna", "to": "ccna-2026"}'
import httpx
httpx.post(
    "http://<host>:8081/api/v1/folders/rename",
    json={"from": "ccna", "to": "ccna-2026"},
).raise_for_status()

There is no folder object to create or delete. A folder exists while a lab claims it and stops existing when the last one leaves — so there is no way to end up with an empty folder you have to tidy, and no way to lose a lab by deleting the folder around it.

Clone

Clone copies the topology — nodes, links, networks, geometry, config sets — under a new name. It does not copy running state; the clone starts stopped.

This is how you build a class: get one lab right, clone it per student, hand out the copies.

It is also the safe way to try something. Clone, break the clone, keep the original.

Lock

Lock makes a lab read-only. Nodes still start, stop and console, but the topology cannot be edited and the lab cannot be deleted.

Use it on the lab you spent an afternoon getting right, or on anything a class is working from. Unlocking is one click — the point is not security, it is not deleting the wrong thing at 9am.

Config sets

A config set is a named collection of startup configuration, one entry per node. A lab can hold several.

Capturereads what the running nodes have now and saves it under a name
Applypushes a set back onto the nodes
Editchange a node's text by hand

The workflow that makes this worth having:

  1. Build the lab, configure it properly, capture as working.
  2. Break something deliberately, capture as broken-ospf.
  3. Hand out the lab with broken-ospf applied. When someone is stuck, apply working and let them read it.

Resetting a class to a known state is one apply rather than twenty consoles.

Import

FormatWhat it is
labtris-lab-v1Labtris's own export. Round-trips exactly
.unlLegacy topology XML (as produced by older emulators)
.clab.ymlcontainerlab

The two foreign formats go through one parser, so they cannot disagree about what a template means.

Anything that cannot be carried over becomes a labelled placeholder and a warning, not a rejection. A 17-node topology that arrives with 14 real nodes and 3 placeholders is worth more than an error message — the warnings say exactly what did not map, and the placeholders hold the shape and the wiring so you can substitute images yourself.

Export

Export writes the lab as JSON. It is the topology, not the images — a lab file is kilobytes and moves by email.

Backup

Settings → Backup downloads every lab plus the instance's settings as one .tar.gz.

Configuration only. Images are deliberately not in it: they are tens of gigabytes, they are yours rather than ours to redistribute, and a backup you cannot email is one nobody takes.

Restore is additive. Labs are created alongside what is already here rather than replacing it — skip (the default) leaves labs this instance already has alone, clone brings them in under a suffixed name, which is what you want when merging two instances. A restore that silently deletes is not a restore.

Google Drive is wired up if you want backups off the box; Settings → Backup walks through linking an account, pushing, and pulling one back.

Next

Users and access — accounts, the two roles, and what ownership actually protects.

← ObservingUsers and access →