Don’t eat my data - 30+ years of storage war stories
summary
Thirty-five years of storage war stories lands on concrete fixes: using DDrescue, TestDisk and PhotoRec to recover a manuscript and a degraded RAID 5, forcing-assembling an 8–disk array to salvage 8–9 TB in hours instead of days, and switching out buggy Marvel PCIe SATA HBAs to stop DMA errors. Covers using mdadm and kernel DM to reassemble arrays, benchmarking with hdparm/fio, running erasure coding like 8+3 across 16 disks, and how plaster dust and humidity killed an entire room of drives.
AI-written summary · about
chapters (6)
- 0:00Introduction
- 1:08My journey with digital data
- 6:54Can I fix this
- 18:00Clusters
- 25:18Dust and humidity
- 30:39What next
related talks
53mOvercome Blast Radius Anxiety With Ultrafast Rebuilds and Performance for High-capacity SSDs
45mBig Data and SSDs
30mCluster storage systems gotta have HeART...
19mEvaluating File System Reliability on Solid State Drives
15mRestoring FDB after Catastrophic Accidents
29mSaving 6 Million Files Under Total Hetzner Conditions
01h 00mA Brief History of Data Storage
34mWar Stories: Fighting the PostreSQL Support Trenches
