Element 9.21
Testing and repairing reused computers
- Sheet
- 9.21
- Chapter
- 9 Manufacturing and repair
- State
- Drafted
- Release
- 16 November 2026
- Editor
- Contributed by Praecise as open research
- Reviewers
- To be named at review
- Safety
- S2, as the chapter
- Licence
- CC BY-SA 4.0
- Resources
- 10
The element
- Requirement
Every reused machine is tested before it carries the community's services and again when it misbehaves; faulty memory, disks and power supplies are found and replaced, and firmware is kept current.
- How to tell
- Memory has passed a full run of a standalone tester such as Memtest86+, which tests more thoroughly than the firmware's own check and runs until stopped (Memtest86+).
- Every disk's self-monitoring data has been read and is watched continuously; smartmontools gives early warning of degradation for SATA, SAS and NVMe disks (smartmontools).
- Firmware is at the latest version the maker provides; in the European Union the maker must provide it for servers (Servers and data storage products: EU ecodesign requirements (Regulation (EU) 2019/424)).
- A written log records each test, fault and part replaced.
- Options
- Memory. Memtest86+ boots from USB, CD or the network on BIOS or UEFI machines, includes bus stress and bit fade tests and an optional Rowhammer test, and checks each module's identity data (Memtest86+). ECC memory is highly recommended for ZFS because without it rare bit flips can go undetected (OpenZFS FAQ).
- Disks. smartctl reads a disk's state on demand; smartd watches it continuously (smartmontools). For hosts running virtual machines, Proxmox recommends SSDs with power-loss protection and discourages consumer SSDs (Proxmox VE System Requirements).
- Remote hardware management. Servers with a management controller that supports Redfish, the DMTF standard for remote hardware management, can report and be managed over the network (Redfish Developer Hub).
- Power supplies. A failed supply is replaced as a whole unit, never opened: it holds mains voltages. A replacement can be chosen by its 80 PLUS level (80 PLUS PSU ratings explained).
- Reusing parts across machines. Memory and storage are the largest share of embodied carbon in the computing systems studied (Sustainability in HPC: Vision and Opportunities), so moving good modules and disks from a retired machine to a running one keeps that value in use.
- Repairability in general. See the library entries on repair already listed for domain 9, such as Rules promoting the repair of goods: Directive (EU) 2024/1799 and EN 45554: assessment of the ability to repair, reuse and upgrade energy-related products.
- Settled, open, contested
Settled. Memory and disk faults are common in used machines and are found by standard free tools (Memtest86+, smartmontools). Open. Failure rates for reused enterprise parts at community scale are not documented in the sources opened. Contested. Whether ECC memory is required: OpenZFS calls it highly recommended, not required (OpenZFS FAQ); many reused desktops cannot take it.
The four stages
- Learn: what each test finds (Memtest86+, smartmontools).
- Design: decide what passes and what is retired; keep a stock of the commonest spare parts.
- Build: test every machine on arrival and record the result (Memtest86+, smartmontools).
- Deploy: run continuous disk monitoring, apply firmware updates on a schedule (Servers and data storage products: EU ecodesign requirements (Regulation (EU) 2019/424)), and log every fault.
Resources at each stage
1 Learn
Understand
2 Design
Size and choose
- 80 PLUS PSU ratings explained Standard, S0
- EN 45554: assessment of the ability to repair, reuse and upgrade energy-related products Standard, S0
- Servers and data storage products: EU ecodesign requirements (Regulation (EU) 2019/424) Standard, S0
- OpenZFS FAQ Guide, S1
- Proxmox VE System Requirements Guide, S0
4 Deploy
Install, run and measure
- Redfish Developer Hub Standard, S0
- Memtest86+ Tool, S0
- smartmontools Tool, S0
What changes between 30 adults and 300 people
At 30 adults one person tests machines as they arrive, with a spare disk and a spare power supply on the shelf. At 300 people testing is a routine with a bench, a log and two trained people, machines are standardised on as few models as possible, and parts are taken from retired machines of the same model.
Reference designs
Open for contributions Designs with drawings, specifications and a bill of materials. Add to it
Parts list
Every resource that supports this element. Those marked External carry the label “External resource. Not made or verified by Public Protocol”.
| Resource | Type | Origin | Maturity | Safety | Licence, as stated at its source |
|---|---|---|---|---|---|
| Redfish Developer Hub | Standard | External | Established | S0 | Not stated on the page opened |
| 80 PLUS PSU ratings explained | Standard | External | Established | S0 | Not stated on the page opened |
| EN 45554: assessment of the ability to repair, reuse and upgrade energy-related products | Standard | External | Established | S0 | Not stated on the page opened |
| Servers and data storage products: EU ecodesign requirements (Regulation (EU) 2019/424) | Standard | External | Established | S0 | Not stated on the page opened |
| Rules promoting the repair of goods: Directive (EU) 2024/1799 | Standard | External | Established | S1 | Not stated on the page opened |
| Memtest86+ | Tool | External | Established | S0 | GNU GPL v2 |
| OpenZFS FAQ | Guide | External | Established | S1 | Not stated on the page opened |
| Proxmox VE System Requirements | Guide | External | Established | S0 | Not stated on the page opened |
| smartmontools | Tool | External | Established | S0 | GNU General Public License, version 2 |
| Sustainability in HPC: Vision and Opportunities | Paper | External | Concept | S0 | Not stated on the page opened |
Depends on, and completes
Depends on 13.1.
It is one of 34 elements that complete Manufacturing and repair.
The standing questions
Who can afford this: the test tools are free; spare parts cost money. Who does the work: two people trained to test and swap parts. Who decides: when a machine is retired rather than repaired is set by a written rule the community agrees (15.2).
Environments
No source found on how testing or repair changes on islands, in arctic cold, in hot deserts, in tropical humid forest, in temperate forest, at high altitude or on flood plains. Where spare parts take weeks to arrive, a larger stock of spares follows from the same logic.
Caution. Power supplies and the inside of a running computer carry mains voltage. Work inside a machine only with it unplugged; never open a power supply.
Safety
State and evidence
Drafted
An editor has written the element's text from the resources listed under it, citing them. It is checked when it is reviewed.
Evidence from a real place or system is recorded here, with what did not work. What the states mean
History
- Drafted in the release of 16 November 2026.