Server problems

This article describes opportunities for the error analysis during a server failure.

Server cannot be started

The server cannot be started successfully.

Verification of information on status-LEDs and acustic warn signals

Verify the following things:

Do certain LEDs shine or blink on the server?
Are there acustic warn signals (or one of the components such as the RAID-controller)?

Consult the guides (server, mainboards, components). The meanings of LED-codes as well as acustic signals are documented there. If the errror cannot be defined, perform the steps listed below.

Isolation of servers and connection of dedicated monitor and keyboard

Perform the following steps:

Remove all cables from the server (power cable, network cable, monitor cable, keyboard cable, mouse cable, etc.).
Wait for 10 seconds.
Now reconnect all power cables.
Now connect a dedicated monitor and keyboard (no KVM-switch). Verify if monitor and keyboard function on another computer.
Now start the server.

If the problem continues to exist, perform the following steps.

Testing servers in minimal configuration

Perform the following steps:

Remove all cables from the server (power cable, network cable, monitor cable, keyboard cable, mouse cable, etc.).
Open the server case.
Disconnect all optional cables that are not required for minimal operations from the mainboard (for example Onboard-SATA cable).
Remove all PCI/PCIe cards.
Remove as many RAM memory modules as possible and leave only the minimum number required for operation in the motherboard.
Close the server case
Now reconnect all power cables, the monitor, and the keyboard
Now, start the server.

You can optionally expand the second CPU for Dual-CPU systems in another step. If the problem continues to exist, perform the following steps.

Switching components

Switch the following components systematically in the listed order. When replacing components, always ensure that all cables (power cable, network cable, monitor cable, keyboard cable, mouse cable, etc.) are disconnected from the server beforehand. After replacing a component, always test whether the server is working properly.

working memory
power supply
CPU
mainboard

Running server, but occuring problems during operation

Network unreachable

Install the current network driver
Isolate the server from the network and try directly via crossover cable to connect to another network device. This ensures that no additional components in your network (switch, firewall, etc.) interfere with operation.

Raid performance is slow

Ensure that you use drivers and firmware from the same manufacturer's code set.
Verify the Raid state. If the Raid state is in "initializing" or "degraded", you have weak performance until the Raid is in optimal condition again.
Verify the settings of the Write Cache. This should be set to "ON" if BBU is available.

Beep sounds

Localise the beep sounds. Components like backplane, controllers or mainboards have the possibility to announce defects via beep sounds.

The Raid is mostly in a suboptimal condition if the server beeps.

Hard drives sporadically fail in the RAID array

In a SCSI system, the terminator must first be replaced or checked.
Make sure that the server is not exposed to strong vibrations (this can be caused, for example, by defect fans)

More information

Author: Thomas-Krenn.AG

At Thomas-Krenn.AG we pay attention to the best possible service. To do justice to this, we have created our Thomas-Krenn Wiki. Here we share our knowledge with you and inform you about basics and news from the IT world. You like our knowledge culture and want to become part of the team? Visit our job offers.

Translator: Alina Ranzinger

Alina has been working at Thomas-Krenn.AG since 2024. After her training as multilingual business assistant, she got her job as assistant of the Product Management and is responsible for the translation of texts and for the organisation of the department.