12 hours later: first lesson.
Had scheduled ZFS storage maintenance - expected 30-45 minutes, took ~2 hours. Everything's back online, but here's why I'm posting about it:
Most hosting companies hide behind vague "planned maintenance" notices. I'm building in public, which means showing the bumps - not just the wins.
Technical details: - ZFS RAID-Z2 pool optimization on 8x 1.2TB enterprise SSDs - RAID-Z2 operations take 3-4x longer than documentation suggests - Next time: budget way more time
Lesson learned: When building in public, transparency matters more than perfection.
Live monitoring (always): https://lightspeedup.com/health.php
Original Show HN: https://news.ycombinator.com/item?id=42059309
Questions: Have you had maintenance take 3-4x longer than expected? How do you handle Day 1 issues?
LightSpeedUp•1h ago
This morning I posted Show HN about LightSpeedUp (Show HN: LightSpeedUp – Hosting with network isolation, 70% cheaper than AWS (lightspeedup.com)1 point by LightSpeedUp 4 hours ago | past).
Tonight: learned that transparency includes admitting when things take longer than planned.
The technical details: - Dell PowerEdge R730, 8x 1.2TB enterprise SSDs in ZFS RAID-Z2 - Ran pool scrub + optimization - ZFS is thorough - checks every block, every parity calculation - Docs said 30-45 min. Reality: 2+ hours.
Why post this? Because most hosting companies would either: 1. Not mention it at all 2. Send a vague "planned maintenance" notice 3. Only talk about it if customers complained
I'm doing something different: complete transparency. When things take longer, I say so. When things break, I'll say so.
That's the core of LightSpeedUp - no secrets, complete transparency, live monitoring at https://lightspeedup.com/health.php
Happy to answer questions about ZFS, RAID-Z2, building in public, or anything else.