I stopped believing in backups the day Sydney deleted a .tfw file.

Sydney was the sysadmin hired to replace the sysadmin who left in 2011.
I was told to train him. Then he became my boss.

One day, he deleted a small file used by TeleForms — something he “thought we didn’t need.”
Within minutes, Data Processing went dark.
TeleForms wouldn’t start. Production halted.

His fix?
Dig through fifty Iron Mountain tapes to find one with that file on it.

Hours later, after sorting, labeling, restoring, and praying, he found it.

The system came back… but the lesson stuck.

See, we already had Compellent Fluid Data — before Dell ever bought them — with Data Instant Replay snapshots sitting right there, waiting to be used.

But Sydney didn’t understand SANs.

I did.

That became my thing — block-based storage, Fibre Channel fabrics, instant recovery.
I even got invited to Sidepath for Compellent training, where the instructor said,

“Wow. You could teach this class.”

Not bragging. Just facts.

Because when you’ve lived through broken systems, you learn the difference between:

• A backup (a copy of what was)
• A recovery plan (a way back to now)

Backups are reactive.
Recovery is architected.

Sydney had backups. I had block storage.

When his tape finally worked, it proved one thing:
Backups don’t save systems. People who understand systems do.

Fast-forward to today — SANs, snapshots, replication, cloud — it doesn’t matter what stack you’re using.
The best “backup” is still someone who knows how to restore sanity.

Final Thoughts

I can already hear the comments, “Snapshots are availability tools, not resilience tools.” Or
“They shorten your RTO, not your RPO.”
“They give you a quick undo, not a clean slate.” Better yet,

“Backups — real, off-array, immutable backups — are your safety net when:

A firmware bug wipes your metadata.

A junior admin formats the wrong volume.

A crypto-worm hits your SAN management interface.

A datacenter flood turns your “five-minute restore” into “five-month insurance claim.”

I’m not saying your love for block-based recovery is wrong; it’s just incomplete.

Snapshots save hours.
Backups save companies.

The trick is knowing when each applies.
Use snapshots for operational recovery.
Use backups for survival.

You don’t need faith in backups — you need discipline in validating them. The day you skip a test restore is the day you start gambling instead of engineering.”

And to that, I say you’re right. Snapshots and replication aren’t the same thing as a backup. But replicating data across arrays in different states mitigates most of those “localized failure” scenarios.

The bigger risk in my world isn’t losing a data center — it’s losing track of where I put something in the first place.

Which is why I sometimes end up with 3–5 copies of the same file scattered across PowerStores, Azure Blob, and my laptop…
because, you know — redundancy. 😅

But seriously: snapshots handle the “oops,” backups handle the “oh no.” I just happen to spend a lot more time fixing the oops.

#Storage #BCDR #SysAdminLife #VMware #ITLeadership #RecoveryNotBackups