Note: This is a fictionalized account. If it sounds exactly like your company, that’s just because dysfunction has a standard operating procedure.

The Disk Explosion Holiday Special

Most people spend Easter weekend hiding eggs.
We spent ours hiding free disk space.

I was on vacation when it started. The data warehouse had been sluggish for months. Queries ran long, ETLs limped through the night, dashboards arrived somewhere between breakfast and lunch. We were already nearly out of disk space, but instead of fixing the cause, someone decided the solution was obvious: rebuild everything.

And by “everything,” I mean:

  • Every index on the warehouse server.
  • Fill factors added where none existed before.
  • Heaps rebuilt, because why not shuffle unordered data into a different order? It does nothing for performance, but it eats CPU and disk I/O so it looks like progress.
  • PAD_INDEX turned on, because the only thing better than wasting space is wasting it deliberately.

Remote Warnings, Ignored

I was away, but still reachable enough to ruin the party. I calmly explained:

“This will not fix the performance problem. And since we are already low on disk, this will likely cause everything to stop working.”

Their counter-argument was elegant:

  • Buy emergency disk space.
  • Do the rebuilds anyway.

Because nothing says “we understand performance” like expanding SAN storage to make room for indexes that were never the problem.

The Predictable Miracle

The outcome was exactly what anyone could have predicted:

  • Disk consumption ballooned like a parade float.
  • Queries did not run faster.
  • ETLs did not finish earlier.
  • The only measurable change was that our MDFs now carried extra frosting.

From an emotional standpoint, progress had been made. From a technical standpoint, we had successfully upgraded from:

  • Before: nearly out of disk, bad performance.
  • After: still nearly out of disk, bad performance, but rebuilt.

Easter Gifts

Meanwhile, my vacation photos feature scenic mountains and subtle background SAN alerts.

Some people get lilies and hymns.
I got tempdb gasping for air and PAD_INDEX as an Easter gift.

And the kicker: the team called it “progress.”

Disclaimer: Any resemblance to actual companies is purely coincidental. Because, surely, no real organization would rebuild heaps, add PAD_INDEX, and buy more SAN space in lieu of actual performance tuning.