A disaster recovery facility is a building whose entire reason for existing is a day you hope never comes. When the primary data center floods, burns, or drops off the grid, this is the site that keeps a bank's transactions clearing or a hospital network's records online. That mission changes how you build it. On a normal commercial job, a two-day slip on cooling commissioning is a scheduling headache. On a DR build, a system that "mostly works" is a system that will fail the exact hour it's called on. The whole job is about proving redundancy, not just installing it.
I've run enough mission-critical work to tell you the schedule is where these jobs are won or lost. Not because DR facilities use exotic materials, but because everything has to be tested, retested, and witnessed before anyone will sign off on activation. The sequencing between concrete, power, cooling, and commissioning is unforgiving. Here's how the real work breaks down and where a look-ahead schedule earns its keep.
Why the Schedule Behaves Differently on a DR Job
On most projects, substantial completion is the finish line. On a disaster recovery build, substantial completion is roughly the halfway point of the part that matters — commissioning. You're not done when the building is done. You're done when an independent commissioning agent (CxA) has watched every redundant system fail over under load and come back clean.
That reality flows backward through the whole schedule. Every trade needs to understand that their finish date isn't when they leave the floor — it's when their work has been energized, tested, and integrated with everyone else's. A short-interval look-ahead is where that gets enforced, because the three-week window is close enough to reality that you can actually see the collisions coming: the electrician who wants to close out gear the same week the CxA needs it open for load-bank testing, the mechanical crew ready to charge the CRAC units before the generators can carry them.
Site Work: The Part Everyone Underestimates
DR sites are deliberately placed away from the primary facility and, ideally, out of the same flood plain, seismic zone, and utility grid. That geographic separation is the whole point, and it usually means you're building somewhere with thin infrastructure. Long utility runs, a new access road, sometimes a well and septic because there's no municipal service within miles.
The gotcha that bites people: the underground and the utility coordination are almost always the real critical path, not the building. You can frame a data hall fast. You cannot make a utility company energize a new feeder faster than their queue allows, and you cannot pour a fuel-tank containment slab until the environmental review clears. Put those long-lead, third-party-dependent items on your look-ahead early and keep them visible. The buffer you want between "trench open" and "utility inspection" is not one day — build in enough slack that a failed inspection or a rain week doesn't cascade into the whole electrical rough-in.
Power Redundancy Is the Spine of the Whole Job
Everything in a DR facility hangs off the power topology, and the topology is the first thing to nail down. Most serious facilities target N+1 at minimum — one redundant unit for every N needed — and the higher tiers push toward 2N, meaning two completely independent paths from utility to the rack. Your schedule has to respect that the redundant path is not an afterthought; it's a parallel build that has to be tested independently.
The typical power sequence, and the coordination traps in each step:
- Utility feed and switchgear. Long lead on the gear itself — often 30 to 50-plus weeks depending on the market. If the switchgear submittal isn't approved in the first month, it will drive your finish date, full stop. Track the submittal and the fabrication milestones on the look-ahead like they're on-site activities, because they gate everything downstream.
- Generators and fuel. Standby generators need their pads, exhaust, and fuel supply all coordinated. The classic miss is treating the fuel system as a separate small package — it's not. A generator with no commissioned fuel polishing and transfer system is a very expensive paperweight.
- UPS and batteries. These bridge the seconds between a utility drop and the generators picking up the load. Battery rooms have their own ventilation and spill-containment requirements that have to be built before the batteries land.
- Distribution, PDUs, and busway. This is where you finally reach the racks, and where the trade congestion peaks.
The single most useful thing a weekly work plan does here is force the electrical, mechanical, and controls trades to agree on the order of energization. You cannot load-bank test the generators until they're fueled. You can't prove the UPS until the generators can carry the downstream load. You can't do integrated systems testing until both are live. Lay that dependency chain out in a look-ahead and the fights happen in the trailer over coffee instead of on the floor at 2 a.m. during a witnessed test.
Cooling: Sequenced to Power, Not to Drywall
Redundant cooling — CRAC or CRAH units, chilled water loops, sometimes economizers — protects the equipment from cooking itself during a long recovery event. The scheduling lesson here is that cooling commissioning is chained to power availability, not to building completion. You can have the mechanical rooms buttoned up and the units set, but you can't run a real cooling performance test until you have stable power and a heat load to reject.
Two things worth building into the plan. First, get the chilled water piping flushed, pressure-tested, and treated well before you need it — a contaminated loop found during commissioning is a schedule killer because you're draining, cleaning, and refilling systems that touch everything. Second, coordinate the load banks. To prove cooling under realistic conditions you simulate the server heat load with resistive load banks, and those load banks need power, floor space, and a testing window that doesn't collide with the electricians closing out gear. That collision is exactly the kind of thing a three-week look-ahead surfaces while there's still time to move it.
Diverse Telecom: The Redundancy People Forget to Verify
A DR site with a single fiber path is not a DR site. You need physically diverse routes — ideally different carriers entering the building through different points, so a backhoe through one conduit doesn't take you dark. The construction piece is straightforward: two entrance rooms, two conduit paths, meet-me rooms. The scheduling piece is that carriers move on their own clock, and their provisioning lead times are long and outside your control.
Get the carrier orders placed and the entrance facilities roughed in early, and — this is the part crews skip — actually verify the diversity. More than once I've seen two "diverse" carriers turn out to share a single pole line a mile down the road. Put a diversity-verification milestone on the schedule so it's someone's explicit job, not an assumption.
Physical Security and the Operations Center
Access control, mantraps, surveillance, and intrusion detection all have to be commissioned and integrated, and security cabling likes to run in the same congested pathways as everything else. Sequence it so the security rough-in rides along with the other low-voltage work rather than showing up after the ceilings are closed.
The operations or command center — the room where recovery gets coordinated during an actual event — deserves its own attention on the look-ahead because it's where finishes, technology, and furniture all converge late. Consoles, display walls, and the raised floor or cable management under the desks all have to land in the right order. Nothing exposes a sloppy schedule faster than a video wall crew showing up to a room that isn't painted, with no power at the console positions.
Fuel, Life Safety, and the Support Spaces
Extended operation is the whole promise of the facility, and extended operation runs on fuel. Bulk tanks, containment, transfer pumps, day tanks, and fuel polishing all have environmental and fire-code requirements that involve inspectors you don't control. Start those approvals early and keep the fuel system commissioning tied to the generator commissioning — they get proven together or not at all.
Don't neglect the human spaces either. If a real disaster keeps a recovery team on site for days, they need working restrooms, a break area, and sleeping accommodations that actually function. These are easy to leave for last and easy to forget in commissioning, but a DR facility that can't house its own operators during a two-day event isn't finished.
Commissioning: Where the Real Schedule Lives
Integrated systems testing (IST) is the moment of truth. The CxA scripts a series of failure scenarios — pull the utility, watch the UPS carry, watch the generators start and take load, watch cooling hold temperature, watch it all fail back — and everyone stands there and confirms the building does what the drawings promised. On a well-run job, IST is a formality that confirms good work. On a badly sequenced one, it's where you discover the problems you should have caught weeks earlier, now with the owner watching.
Protect the commissioning window ruthlessly. A few habits that keep it honest:
- Treat commissioning as its own phase on the look-ahead with named activities — Level 1 through Level 5 checks, load-bank testing, IST — not a single "commissioning" bar at the end.
- Work the punch and closeout before IST, not after. You don't want a trade opening up gear to fix a defect the week you need everything sealed and stable.
- Build in a re-test buffer. Something always fails the first pass. Assume it, schedule it, and you look like a hero instead of scrambling.
- Capture as-builts as you go. Recovery procedures depend on accurate documentation — when the facility is actually in use, nobody has time to reverse-engineer how the power transfers work.
Where Look-Ahead Scheduling Actually Helps
DR construction is a coordination problem more than a means-and-methods problem. The building isn't hard; keeping a dozen trades and a handful of third parties aligned through a testing-heavy finish is. That's precisely the work a rolling short-interval schedule is built for: pulling the next three to six weeks into a view everyone can see, exposing the dependencies between power, cooling, and commissioning before they become collisions, and giving the foremen a weekly work plan they can actually staff to.
A tool like LookAheadWall fits here because it keeps the plan visual and location-based — you can see which trade owns which data hall which week — and because trade-flow sequencing makes the energize-then-test dependencies explicit instead of living in one superintendent's head. But the tool is only as good as the discipline behind it. The teams that deliver these facilities on time are the ones who treat the look-ahead as a live conversation, updated weekly with the people doing the work, not a wall decoration.
Build the redundancy, then prove it. On a disaster recovery job, the schedule that gets you cleanly through commissioning is the one that respected the testing sequence from day one — and made every trade see it coming.