Local systems
Rebuilding the R640 into a local AI workstation
Building a local AI environment included patient hardware fault isolation, storage correction, and operating-system recovery.
The question
Can the server recover from repeated shutdowns and boot/storage faults into a usable local AI workstation?
How the work was approached
Recorded controlled component isolation, replacement-board work, PMem troubleshooting, SAS cable correction, Linux Mint recovery and software setup.
What the record shows
The log records return to a healthy front-panel/iDRAC state after rebuild, two P4 devices detected, Mint desktop used, LM Studio running, and a transfer page returning HTTP 200 on September 25.
What remains unresolved
Most physical actions are Reported/previously verified summaries. Exact final CPU, memory, PMem and storage inventory is unresolved. October 6 checks could not reach historical endpoints; October 13 remote inventory was rejected by the existing restricted command route. Do not publish IPs or assert current runtime/GPU performance.
What we are trying to achieve
The R640 is intended to provide a stable foundation for local models, retrieval, and bounded computational research. Hardware recovery matters because a research result is only meaningful when the machine’s condition and the execution setting are understood. Present-day specifications still need a dated inventory.
Sources & reading notes
- R640-Recent-Build-and-Repair-Log-2026-09-25.mdChronological Repair and Build Record / Operating System / 2026-10-06 Read-Only Readiness Check Retained project record Source record retained privately.
- CHECKPOINT.mdRestricted R640 inventory route and current unknowns Retained project record Source record retained privately.
This is a dated, source-based research account, not a live machine report. Historical results, engineering preparation, research designs, open review gates, and measured pilots remain distinct.