Dedicated servers · Question & answer

What evidence should I include in a dedicated-server hardware fault ticket?

A useful hardware fault ticket connects the observed symptom to a specific machine, time window and component. Include the impact, recurrence, last known good state and recent changes, with relevant OS and management-controller evidence. Avoid concluding that a part has failed solely from one ambiguous message; state what is observed and what remains uncertain.

IntermediateLinux on a dedicated server; hardware capabilities depend on the server model
Go to the answer

Before you start

  • Authorised access to relevant logs and the provider’s secure support channel.

Instructions

1. Record the incident time and kernel context

Use UTC timestamps to make cross-system correlation easier. A short kernel-log window is more useful than an unexplained full diagnostic dump.

Remote Linux shell
date -u -Is
uname -r
sudo journalctl -k --since "1 hour ago" -p warning -n 100 --no-pager

Adjust the interval to the incident. Preserve exact error codes and device identifiers, while reviewing output for credentials, customer data or unrelated sensitive details before sharing.

2. Add component and reproduction details

Identify the affected serial, bay or PCI address using the appropriate inventory and vendor mapping. Include whether the symptom occurs at idle, under a particular workload or after a recent change, and list diagnostics already performed.

3. Define the requested investigation

State the allowed maintenance window, available console access and desired next step. Keep destructive diagnostics, resets and physical replacement separate from the request to review evidence.

Check the result

  • Support can identify the correct machine and component, correlate the event timeline and understand the permitted investigation without needing passwords in the ticket.

Troubleshooting

The BMC timestamp differs from the OS
Record both clocks and their offset instead of silently changing the evidence.
The fault is intermittent
Provide recurrence times and monitoring samples; a clean snapshot after recovery does not disprove the incident.
A serial number cannot be mapped to a bay
Request provider confirmation before authorising removal.

Official references

Use the documentation for your installed version when a flag, service name or platform feature differs.