A server rarely fails all at once. It warns you first: a fan alarm, a degraded array, a memory error in the log, a power supply light turning amber. The hardware has redundancy built in for exactly that reason, and the right response is to fix the part before the redundancy runs out. This guide covers the faults we are asked about and what hardware repair does and does not include.
Blue Apple repairs servers in Amsterdam. The servers page lists the brands we work with, what typical repairs cost and how to send us a request.
What usually goes wrong with servers
- Power supply failure or redundancy faults. A failed supply is the classic server fault, and a redundant pair means the machine keeps running on one. The risk is the second failing before the first is replaced.
- Fan failure and overheating. A failed fan raises temperatures across the chassis and triggers alarms, throttling or a shutdown.
- Mainboard faults, no POST or boot errors. A server that powers on and never reaches the boot screen, or stops with a hardware error code.
- Memory (RAM) errors. Correctable errors that rise over time, or a DIMM that is flagged as failed in the management log.
- RAID controller and storage backplane faults. A degraded array that will not rebuild, drives that drop in and out, or a controller that is not detected.
- Management controller (BMC/IPMI) faults. A remote management interface such as iDRAC, iLO or IPMI that stops responding while the server itself runs.
Brands and models we look at
These are the server brands we list, and some of the models with their own pages, each with an estimate of what a typical repair costs.
Brands: Dell PowerEdge, HPE ProLiant, Lenovo ThinkSystem, Supermicro, Fujitsu, Cisco UCS, ASUS Server, Gigabyte Server, ASRock Rack, Inspur, IBM, Oracle.
Popular models: Dell PowerEdge R660, Dell PowerEdge R760, Dell PowerEdge T360, Dell PowerEdge R360, HPE ProLiant DL380 Gen11, HPE ProLiant DL360 Gen11, HPE ProLiant ML30 Gen11, Lenovo ThinkSystem SR650 V3.
Repair or replace?
A server is usually worth repairing, because a power supply, fan, DIMM or controller costs a fraction of the machine, and the data and configuration on it are worth far more than the hardware. This service covers hardware repair: operating-system support, software licensing and data recovery are not part of it unless agreed in writing. Before any repair, make sure you have a current backup.
Looking after your server
- Replace a failed redundant part at once. Redundancy only protects you until the second part fails.
- Watch the logs and the fan and temperature readings. The warnings come before the failure.
- Keep the intake clean and the room cool. Dust and heat shorten the life of everything inside.
How a repair works at Blue Apple
The prices for these devices are estimates until we have seen them. Send a request with the model and what is wrong, and we tell you what a repair is likely to cost and how to get the device to us. The final price is confirmed in writing after the diagnosis, before any work starts. You can use the custom repair form, or call 020 239 11 93.
Common questions
Do you recover data from a failed server?
No, not as part of the repair. Hardware repair does not include data recovery unless agreed in writing, so keep a current backup.
Which server brands do you work with?
Dell PowerEdge, HPE ProLiant, Lenovo ThinkSystem, Supermicro, Fujitsu, Cisco UCS and more, on the server page.
In short
Power supplies, fans, memory, RAID controllers and management boards are what fail on a server, and the diagnosis tells you which one you have. See the servers page for the models and repairs, or send us a request. You will find us at Rijnstraat 76-H in Amsterdam, and you can call 020 239 11 93 with a question first.
