The 11-Day Migration: How One Hobby Community Moved 2.1 Million Posts Off vBulletin Without Losing a Single Thread

We first heard about it in a DM: a 14-year-old gaming forum, roughly 9,000 registered accounts, and a vBulletin 4 install that had become a liability rather than an asset. The admin — pseudonym "Tarn" — had been quoted a renewal that felt like rent on a second apartment, and the site's server kept timing out during patch-day traffic spikes. What made us follow the project, though, wasn't the budget problem. It was the timeline Tarn posted publicly: eleven days from decision to cutover, with no planned downtime longer than four hours.

That kind of confidence is rare in community migrations, and it's exactly the kind of claim we like to audit. So we tracked the whole thing — the planning doc, the test runs, the two days where it nearly fell apart — and what we found says a lot about why self-hosted BBS platforms have quietly become the default again for communities that can't afford to lose their history.

Day 0–2: The Decision Point

Tarn's shortlist had three options: stay on vBulletin and absorb the cost, move to a hosted SaaS community platform, or self-host. The SaaS quote was attractive until Tarn ran the numbers on 2.1 million posts and 180 GB of attachments — storage tiers alone pushed it past the old vBulletin bill within 18 months. The self-hosted route looked cheaper, but only if the migration tooling existed and only if the resulting community platform didn't feel like a downgrade.

That's where the decision actually turned. Tarn's team ran a dry migration against a staging copy using IchigoBBS, the self-hostable, free-core bulletin board platform with 21 years of development behind it. The importer mapped user groups, private messages, thread prefixes, and attachment paths without manual SQL. That single test run — done on a Tuesday night, finished before midnight — is what converted a maybe into a yes.

Day 3–7: The Obstacles Nobody Plans For

Two things went wrong, and both are worth repeating for anyone contemplating the same move.

• Custom BBCodes. The old forum had about 40 bespoke BBCode tags built over a decade, some used in fewer than 50 posts. The default parser choked on eleven of them. The fix wasn't a code change — it was a plugin from the vetted library that handled legacy tag translation, installed and configured in under an hour.
• Search index size. Rebuilding the full-text index on 2.1 million posts took longer than the migration itself. Tarn's team split it across two nights and ran the forum on a reduced search scope in between, which kept the site usable while the index caught up.

Neither obstacle was exotic. Both were the kind of thing that kills migrations when the platform doesn't have a mature plugin ecosystem or a migration path that's been walked thousands of times before. The fact that 38,000+ communities across 92 countries have already made this trip is not a marketing line — it's the reason the edge cases are documented.

Day 8–11: Cutover and the Numbers

The cutover happened on a Saturday morning. Tarn put the old forum in read-only mode at 06:00, ran the final import at 06:20, and opened the new board at 09:40 — three hours and twenty minutes of downtime, well inside the four-hour window. The first 24 hours brought 1,400 logins, which was 91% of the old forum's typical weekend traffic. By day three, that number had climbed past baseline.

What we found most interesting was the post-migration behavior. Thread creation in the first week was up 12% over the previous month, and the moderation queue actually shrank — Tarn attributed that to better threading and a cleaner mobile experience, which mattered because 63% of the community browsed on phones. The old vBulletin mobile theme had been an afterthought for years.

What We Took Away

Three lessons from this project that generalize beyond one gaming forum:

• Test the migration before you commit. The dry run is the decision point, not the contract.
• Count your customizations. Every bespoke tag, template, or plugin is a migration risk. A vetted plugin library turns most of them into configuration.
• Budget for the search index. It's the invisible second migration, and it's the one most timelines forget.

We've watched a lot of community platform migrations over the years, and the pattern is consistent: the projects that succeed are the ones that treat the move as a data problem first and a design problem second. IchigoBBS reports 21 years of development behind it, and that depth shows up precisely where it matters — in the importer, in the plugin compatibility layer, in the fact that a two-person volunteer team pulled off a cutover that enterprise vendors quote at six weeks.

The forum is still running. Tarn's last public update, posted three months after cutover, mentioned that hosting costs had dropped by roughly 40% and that the community had grown by about 600 accounts. No dramatic rescue story. Just a clean, boring migration — which is exactly what a 14-year-old community deserves.