Backup & Recovery, Tested Before You Need It.
If it broke tonight, when would you be working again? Backups are the easy part. We build for the answer to that question, and we prove it while nothing is wrong.
Everyone Sells Backup. Recovery Is the Part That Matters.
A backup is a copy of your data. Recovery is the process that turns that copy back into a working business. They are not the same thing, and nearly every company that has lost data had backups running the entire time.
This Is the Whole Picture Most Businesses Ever See
A row of green ticks and a timestamp. It confirms exactly one thing: a job finished. That is worth knowing, and it is nowhere near enough to plan a bad day around.
What a green tick cannot tell you
- Whether the copy will actually open
- How far back it goes
- How long a full restore would take
- Whether the same stolen login could delete it
- Whether anybody has ever tried
That gap is where the bad weeks live. A business in Eugene does not lose two weeks because nobody made a copy. It loses two weeks because the copy was a month old, or the same stolen login that reached the servers could also reach the backup, or nobody had ever tried a restore and the first attempt happened during the emergency.
The Two Numbers Nobody Gave You
Two questions decide what a backup design has to look like. They are the two numbers every disaster recovery plan is built on. Most businesses have never been asked either one, which is how they end up with protection that does not match what they can actually survive.
How much work can you afford to redo?
Backups happen at points in time. Whatever changed between the last one and the moment things went wrong is gone, and your team recreates it from memory. If backups run once overnight, the honest answer is up to a day of work. Shorten the gap and you shorten the loss. That is a design decision, and it should be a deliberate one.
Known in the trade as the recovery point objective (RPO)
How long can you afford to be down?
This is the clock from the moment something breaks to the moment people are working normally again. It covers noticing the problem, deciding what to restore, moving the data, and checking that what came back is right. It runs longer than most people expect, because having the copy is only the first step of the job.
Known in the trade as the recovery time objective (RTO)
Your Answer Is Not Somebody Else’s
A dental office in Springfield that cannot see patients without its practice software has a very different answer from a manufacturer in Lane County that can run on paper until lunchtime. Nobody can work these out on your behalf, and the design has to follow the answers rather than the other way round. Guessing is how a business ends up paying for protection it does not need, or discovering the gap it cannot survive on the worst morning of the year.
What Actually Gets Protected
Protection is not one setting switched on everywhere. Different parts of a business fail in different ways, so each one gets a design that matches how it breaks and how fast it has to come back.
Servers, as Whole Machines
A server is captured as a full image, so the operating system, the applications, the settings and the data all come back together. Rebuilding from loose files means reinstalling and reconfiguring everything before anybody can work again. An image skips that entire day.
Where restore testing runs
Microsoft 365 and Google Workspace
Mailboxes, calendars, contacts, SharePoint sites, OneDrive libraries and Teams content are backed up separately from the platform that hosts them. Microsoft keeps the service running, and its own service agreement recommends that you keep your own copy of what is in it. It is also the single most commonly missed piece of coverage we find across Oregon, because it feels like somebody else’s responsibility.
Backed up and archived
Laptops and Desktops
The standard design keeps working documents synced off the device rather than sitting on one hard drive, so a lost, stolen or dead laptop costs you a machine and not the work. Hardware is replaceable on a day of notice. A folder nobody copied is not.
The default for most teams
The Machines That Need More
Some workstations hold something you cannot simply reinstall: a design rig, an old line-of-business application, a bench PC nobody documented. Those can take a full image or a document-level backup of their own. It is a deliberate addition, because most machines genuinely do not need it.
Added where it earns its place
A Backup and an Archive Are Not the Same Thing
Both keep copies, and they answer completely different questions. Most businesses own one of them and have never been told the other exists.
A Snapshot of a Moment
A copy of the mailbox or the site exactly as it stood when the job ran. Restore it and you get that moment back. Anything created and then deleted between two runs was never captured, because nothing was watching in between.
A point in time, once a night
A Continuous Record
Every message is captured as it arrives and as it is sent, at the server, before anyone can touch it. A message somebody deletes on purpose is still there. Retention is set to match whatever you are required to keep, from a month to well beyond a decade, and the copy cannot be altered afterwards. Every Microsoft 365 and Google Workspace protection we run is archival for that reason, so retention matches a compliance obligation instead of the platform default of roughly ninety days.
Everything, as it happens
Three Copies, and One Nobody Can Erase
Every backup design gets measured against the same convention. It began as three copies, on two kinds of storage, with one of them off site. Ransomware forced two more additions, and the current version is written 3-2-1-1-0. It is worth knowing, because it gives you a checklist you can hold any provider to, including this one.
Three Copies of Everything
Your live data counts as one. Two more exist somewhere else. The test is not how many copies you can name, it is whether a single event could destroy more than one of them at the same time. Two copies in the same building are closer to one copy than to two.
Two Different Kinds of Storage
Copies that share a failure mode are one copy wearing a disguise. The standard design puts a copy in the cloud. Where a business wants faster local recovery, a second copy lands on a device in the building, which is genuinely different hardware in a genuinely different place. The cloud copy comes first, because on its own it covers off site, immutable and verified. The local device is an addition rather than a foundation, and it earns its place when a business cannot wait on a download to start working again. For a Eugene office with a few hundred gigabytes that wait is an afternoon. For one with several terabytes it is precisely why the local copy exists.
One Copy Off Site
The cloud copy is the off-site one. A fire, a flood, or a break-in that takes the building does not take the backup along with it. This is the copy that separates a hard week from a business that never reopens, and it is the part of the design people mean when they talk about disaster recovery.
One Copy That Cannot Be Changed
The cloud copy is written so it cannot be altered or deleted afterwards, including by an administrator account with the right password. This matters more than it used to. Ransomware now looks for the backups first and encrypts them before touching anything you would notice. A local device on the network is fast, and it is not immune to that. The immutable copy is the one that is.
Zero Doubt That It Works
A backup nobody has ever restored is a theory. Servers are started from their own backups on a schedule and photographed running. That is the last digit, it is the one most providers quietly skip, and it is the whole of the next section.
A Backup Nobody Has Restored Is a Theory
This is the part of the job most providers describe and almost none of them evidence. A server is started from its own backup on a schedule, and the screen it comes up on is captured while it runs. Not a report that the job finished. The machine, running, with a picture to prove it.
What This Covers
Servers protected as full images are started from their backups on a schedule and photographed running. It happens on its own, without anybody remembering to arrange it, and it is part of the service rather than an extra. If a test fails, that is a problem discovered on an ordinary weeknight instead of during an emergency.
What It Does Not Cover
Boot testing applies to server images, because a mailbox has no screen to come up on. Files, mailboxes and cloud data are verified a different way: by periodically restoring real items by hand and confirming they open. Less elegant, entirely manual, and the only honest way to know that half works.
Ask any provider in Eugene or Springfield when your last restore was tested and what it produced. A date and something you can actually look at is an answer. A dashboard full of green ticks is the thing that gets mistaken for one.
The Product Is the Easy Half
Any provider can license the same platform and switch the same settings on. What decides whether a bad morning stays a morning is the part that is not in the software: who sequences the restore, who decides how far back to go, and who picks up when you call.
The Restore Is Sequenced
When several things have to come back at once, the order decides how the day goes. The system you cannot trade without comes up first and the rest follows behind it. That order comes from knowing how your business actually runs, not from a default in a console.
Somebody Decides How Far Back
Restore too recent and you bring the problem back with the data. Go too far back and you have thrown away work that was perfectly fine. Somebody who knows your environment makes that call, instead of it landing on you at six in the morning.
The Person Who Built It Answers
The design, the testing and the phone call are the same person. Nothing has to be explained twice, and there is no queue between you and somebody who already knows exactly how your setup is put together.
Somebody Can Come to the Building
Some failures need a person standing in front of the hardware, not a remote session. Being in Lane County means that is a drive across town rather than a ticket waiting on a dispatch queue.
When Was the Last Time Anyone Tested a Restore?
Not whether the backups ran. Whether somebody has taken a copy, put it back, and watched it work. If you cannot name a date, that is worth a conversation. We will tell you what is actually protected today, what is not, and what it would take to know the difference.
Questions We Hear a Lot
The things people actually want to know before they change how their data is protected.
Something not covered here?
Ask it directly and you will get a straight answer, not a sales call.
Get in Touch →What is the difference between a backup and a recovery?
A backup is a copy of your data. Recovery is the process of turning that copy back into a working business, which involves deciding what to restore, moving it, and checking it came back correctly. Almost every company that has lost data had backups running at the time. What they did not have was a tested way to get back.
What do RTO and RPO mean?
They are the two numbers that define a backup design. The recovery point objective (RPO) is how much work you would have to redo, measured by the gap between backups. The recovery time objective (RTO) is how long you would be down, measured from the moment something breaks to the moment people are working normally again.
Is Microsoft 365 backed up already?
Not in the way most people assume. Microsoft keeps the service running and protects its own infrastructure, but the data inside your tenant remains your responsibility, and the Microsoft Services Agreement recommends you keep your own copy. The recycle bin holds deleted items for roughly ninety days, which is a retention window rather than a backup.
Does Google Workspace need its own backup?
Yes, for the same reason as Microsoft 365. Google protects the platform, not your decisions about the data in it. A file deleted in error, a mailbox emptied by somebody leaving, or an account compromised and cleaned out are all your side of the line. We back up and archive both platforms the same way.
What is the 3-2-1 backup rule?
Three copies of your data, on two different kinds of storage, with one of them off site. The modern version is written 3-2-1-1-0, adding one copy that cannot be altered and zero errors on a verified restore. It is a useful checklist because you can hold any provider to it, including us.
Can ransomware encrypt our backups?
It tries to, and this is why immutability matters. Modern ransomware looks for backups first and encrypts or deletes them before touching anything you would notice, often using stolen administrator credentials. A backup on a device on the same network is reachable. The cloud copy we keep is written so it cannot be altered or deleted afterwards, including by an administrator.
How often should a backup be tested?
Regularly, and by starting the backup rather than reading a report about it. Servers protected as full images are booted from their own backups on a schedule and the screen is captured as evidence. Files, mailboxes and cloud data are checked by restoring real items by hand, because there is no screen for a mailbox to boot to.
Do you provide backup and recovery for businesses in Eugene and Springfield?
Yes. PathWise IT works with small businesses across Eugene, Springfield and the wider Lane County area in Oregon. Being local matters when a failure needs somebody standing in front of the hardware rather than a remote session and a dispatch queue.
Let’s Find Out If You Can Recover.
If you cannot name the date of your last tested restore, that is the place to start. We will look at what is protected today, what is not, and what it would take to know the difference. You will know where you stand by the end of the conversation.