From Wildfire to Failover: A Higher Ed Practitioner’s DR Story

This blog shares Julio’s first‑hand experience running infrastructure at Mount St. Mary’s University during the LA wildfires and what happened when one of their campuses was threatened. Nikita Bhuma from Rubrik adds practitioner notes and context for others designing DR and Microsoft 365 protection in higher education industry.

I’ve been the Infrastructure Administrator at Mount St. Mary’s University for over 13 years, and in that time I’ve seen a lot of backup tools promise simplicity and then turn into a part‑time job to manage.

Before Rubrik, we were on a Dell backup solution that was fine on paper, but too convoluted for a small higher ed team with no dedicated FTE just for backup. It did the job, but it constantly needed care and feeding.

When we evaluated Rubrik back in 2017, what immediately clicked for me was how appliance‑like it felt. My joke at the time was: “It’s like a toaster – once it works, it works.” That sounded like a throwaway line back then. It stopped being a joke during the LA fires.

Our infrastructure is split across two campuses, each running a different hypervisor platform. For data protection, we use two Rubrik appliances, covering both our on-premises Virtual Machines (VMs) and critical Microsoft 365 workloads.

Nothing exotic, but enough complexity that DR can’t just be “we’ll figure it out on the day.” Cross‑site replication and clear runbooks became table stakes for us once we moved to this dual‑campus model.

Earlier last year, the Los Angeles wildfires stopped being something we watched on the news and became very real for one of our campuses.

• The fire line was close enough that access to the site was restricted.    
• Power was unstable and our backup generator was burning through fuel.    
• We were staring at a likely full site shutdown with no safe way to “go in and fix it.”    

This is where all the unglamorous prep work mattered. We didn’t suddenly invent a new DR strategy; we used what we already had with Rubrik and our secondary site.

We failed over to the other campus and ran from there. No heroics, no fancy scripting showcase. It worked the way we’d planned it on paper, which, in that moment, was exactly what we needed. The “toaster” did its job.

From a practitioner’s point of view, that was the real return on investment: when everything else around you is on fire (literally), your data protection stack behaves predictably.

Dual‑site higher ed environments tend to accumulate complexity over time – mixed hypervisors, different storage, and a growing M365 footprint. A consistent protection layer across those pieces is usually the first step toward practical DR

The wildfires were the loud problem. Microsoft 365 governance is the quiet one that never really goes away.

As Teams and SharePoint took off on campus, people started treating them like personal file shares and long‑term archives. In higher ed, that creates a tension:

• Faculty and students want freedom to collaborate.    
• The institution still needs to meet compliance, retention, and legal hold requirements.    

Rubrik’s M365 protection gave us a way to walk that line:

• We can set sensible retention and protection on SharePoint sites, mailboxes, and Teams.    
• We have granular restore when (not if) something important gets deleted.    

Real examples we deal with:

• A professor accidentally deletes course content mid‑semester.    
• We need to recover a specific email thread or SharePoint folder for an audit.    

Being able to restore just that one site, that one folder, or that one mailbox without a giant restore operation has saved us hours each time.

If you’re running infrastructure in a similar environment, here’s what I’d emphasise:

• Design assuming you can’t touch the site.
  During a real‑world event, you may not be allowed anywhere near your data center. Cross‑site replication and a documented failover path are not “nice to have” in disaster‑prone regions.
• Test recovery, not just backups.
  We had run DR on paper and in smaller tests. The wildfire was the first full‑scale validation. Make time to rehearse: who does what, in what order, and how you’ll know it worked.
• Treat Microsoft 365 as critical infrastructure.
  For us, SharePoint and Teams are part of how the university functions day to day. Protect them with the same mindset as your VMs: clear policies, clear ownership, and the ability to restore at a very granular level.
• Keep operations simple under stress.
  When you’re watching generator fuel levels and fire maps, you don’t want a DR workflow that requires ten custom scripts and three people to remember the magic sequence. Aim for predictable, repeatable steps.

After 13 years in this role and 7+ years on Rubrik, reliability matters more to me than anything else. During the LA fires, while access roads were blocked and our generator was running low, our data and services kept running from the backup site the way we had planned. At that moment, the old “it’s like a toaster” line felt pretty accurate.