Stories

Evidence from people who opened the chassis with us.

These notes reference specific courses, constraints, and outcomes. Voices are lightly edited for clarity; names used with permission.

“During Unstable Rig Diagnostics we finally timed our thermal probes instead of guessing from exhaust smell. The afternoon drill with a deliberately loose PSU cable felt unfair until the same fault showed up on our own row two weeks later.”

— Anong W., shift supervisor · Unstable Rig Diagnostics

“Hashboard workshop taught us to refuse warranty claims without photos of the connector edge. We still scrap more boards than I’d like — the class was honest about that — but we stop shipping units that only needed a reseat.”

— Pedro L., parts desk · Failed Hashboard Recovery

“The network lab’s intentional DHCP storm was chaotic. I left with a checklist taped above our patch panel. Pool latency alerts dropped once we fixed a duplex mismatch we had ignored for months.”

— Soraya K., network caretaker · Network Fault Tracing

“Private mentoring at the Bang Bua Thong bench helped our new hire shadow a full triage. Travel from Rayong was long; the half-day still beat another week of random part swapping.”

— H. Tanawat, owner-operator · Mentoring session

Extended case

Weekend instability on a twelve-rack GPU shed

A Chonburi team arrived after three weekends of random reboots. In class they logged inlet temperatures, then discovered their “unstable” shelf shared a UPS with a compressor that kicked every twenty minutes. The course did not sell them a new UPS — it gave them a logging habit that made the compressor correlation obvious on Monday.

They returned for the network lab after fixing power, and found leftover static IPs colliding when miners were hot-swapped. Two courses, one quieter shed.

Tell us what keeps failing