Lineage 2 Server Roadmap: From Idea to Launch in 2026

A lineage 2 server roadmap is mostly a schedule, not a stack of tutorials. The technical parts are documented and take two weekends. What kills projects is doing them in the wrong order, announcing before the client works, or opening to 200 people with a bug you could have found with five. This roadmap runs from the first decision to the first 30 days live, and every stage ends with a gate. If you cannot pass the gate, you do not move on.
Stage 0: Decide What You Are Building
Before you download anything, write one page. Chronicle, rates, target population, PvE or PvP focus, and how you expect the server to pay for itself. That page is the reference you will use every time a decision is unclear, and it takes an evening.
Be specific. "Interlude, x50, PvP, 150 players at launch, donation shop with cosmetics only" is a plan. "A cool L2 server" is not. If you cannot name the chronicle and the rate, you are not ready to install, because both choices decide which files you use, which client you need, and who your audience is.
Gate: a written one-page spec you would be comfortable showing a stranger.

Stage 1: Build the Stack Locally
Pick your files before you pick a host. L2JMobius covers dozens of chronicles and is the practical choice if you want High Five, Classic or anything modern. aCis is Interlude only and aims at retail accuracy, which is why so many Interlude servers start there.
Install the JDK the pack asks for, not the newest one. Set up MariaDB or MySQL, import the pack's SQL, build with the tool the project ships, then start the login server and the game server as separate processes. Expect one weekend for the first successful build and another for a clean rebuild on a fresh machine, because that second run is what you will repeat on the VPS.
Gate: you log in on 127.0.0.1, create a character and walk around.
Stage 2: Configure the World Before Anyone Sees It
This is the stage owners rush and regret. Rates, drop and spoil tables, spawn density, teleport and buffer NPCs, offline trade, safe zones, subclass and Noblesse requirements, and whatever custom content you promised. Change one thing at a time and write down what you changed, because a server that behaves unpredictably in week two is usually a config file nobody remembers editing.
Two rules save months. First, do not mix a custom feature with a balance change in the same test. Second, keep a clean copy of the original config directory, because "restore defaults" is the fastest fix in this scene.
Gate: you play for an hour on a test client and hit no blocker.

Stage 3: The Client and the Launcher
Players do not see your server, they see your client. The client version must match the chronicle and the protocol your pack speaks, L2.ini has to point at your IP, and the patch system has to deliver updates without asking anyone to reinstall 6 GB.
You can run a bare client and a manual patch for a closed beta. You cannot run one for a public launch. A launcher with auto-update is the difference between a player who stays and a player who gives up at the download screen.
Gate: a person who is not you installs the client and reaches the character screen without asking a single question.
Stage 4: Website, Discord and the Listing
You need three things online before you invite anyone. An account panel that registers players and, if you use vote rewards, tracks them. A Discord where announcements and support live. And a listing where players can actually find you, with your chronicle, rates and opening date filled in correctly.
List the server on L2Calendar through the free listing form as soon as your opening date is fixed, because the calendar is what players browse when they plan the next few weeks. VIP placement can wait until you see whether the free listing converts.
Gate: a friend registers on a phone, logs in, and joins your Discord from a link on the site.
Stage 5: Closed Beta With Strangers
Ten to twenty people who are not your friends will find bugs your test client never showed. Give them two weeks, tell them the world will be wiped, and ask for one thing: a written report of anything that felt broken, unclear or unfair.
Run a load test while they are there. Thirty clients on the same machine, a mass PvP in one zone, and one castle siege if the chronicle has sieges. Watch memory, watch the login queue, and watch whether the database keeps up. This is the cheapest moment in the project to discover that 4 GB of heap is not enough.
Gate: two weeks with no critical bug and one load test that did not fall over.
Stage 6: Launch Week
Announce one to two weeks ahead, no more. A date announced a month early gives people time to forget, and a date announced on the day gives nobody time to plan. Pick the mode: an open beta that carries into launch, or a clean grand opening. Both work, but the second one needs a client that is already patched and a staff roster that is awake.
On launch day, one person does nothing but watch the login server and the Discord. A second person answers tickets. You do not play. The first four hours decide the tone of the next four months.
Gate: server up, patch working, a GM online and visible.
Stage 7: The First 30 Days
Retention, not acquisition. Publish an event calendar for the first month, patch balance issues weekly rather than silently, and answer feedback in public so other players see that reporting something is worth the effort. Players forgive a bug they saw fixed. They do not forgive silence.
Watch two numbers: how many of the launch players return on day seven, and how many are still there on day thirty. Everything you spend on marketing after that should be aimed at the second number, because buying traffic for a server that loses 80% of its players in a week is the most expensive way to feel busy.
Gate: you know your day-seven number and you have changed something because of it.
The Whole Roadmap on One Screen
| Stage | Realistic time | Can you defer it? |
|---|---|---|
| 0. Spec | 1 evening | No |
| 1. Local stack | 2 weekends | No |
| 2. World config | 1 to 3 weeks | Partly, but not rates |
| 3. Client and launcher | 1 weekend to 2 weeks | Launcher only, for a beta |
| 4. Site, Discord, listing | 1 weekend | Panel can start simple |
| 5. Closed beta | 2 weeks | No |
| 6. Launch week | 1 week | No |
| 7. First 30 days | 30 days | It is the job |
Add it up honestly and a first server takes six to ten weeks of evenings. The projects that open in three weeks are not faster, they just skipped stages 2 and 5 and will pay for it in public.
What to Defer and What to Never Skip
Defer the custom content, the second world, the tournament and the mobile app. Defer the paid advertising until you know your retention. None of those decide whether the first month works.
Never skip the spec, the clean rebuild on a fresh machine, the closed beta, or off-machine backups. Those four are what separate a server that had a bad launch from a server that had its last launch.
