Every PowerEdge server includes Dell's Embedded System Diagnostics, also called ePSA (Enhanced Pre-boot System Assessment). It runs before the operating system and sees the hardware directly, so Dell notes it can find problems the OS and online tools miss. It tests memory, CPUs, I/O devices, drives and other components. Run it when a server will not boot, after a hardware alert, after replacing a part, or as routine maintenance.
Before you start
- Diagnostics run outside the OS, so the server is offline for the whole test. Schedule a window.
- You need local console access or the iDRAC9 Virtual Console.
- Lifecycle Controller must be enabled (F2 > iDRAC Settings > Lifecycle Controller).
- Note the Service Tag and the current alerts, and export the logs first (see below).
Step 1 — Save and review the event logs
Dell's KB 000132726 advises clearing the System Event Log before running ePSA. If old unchecked records remain, ePSA may report error 2000-0251. That code points to old log entries, not to a hardware failure.
Warning: Clearing the SEL removes the evidence of past faults. Export or screenshot the SEL and Lifecycle Log first, so you and Dell support still have the history. See iDRAC9 SEL and Lifecycle Log.
Step 2 — Launch ePSA
| Generation | How to start |
|---|---|
| PowerEdge 14G, 15G, 16G | Press F10 during boot, then Hardware Diagnostics > Run Hardware Diagnostics. Or press F11 and choose System Utilities > Launch Diagnostics. |
| PowerEdge 17G (per the same KB) | First install the UEFI ePSA Diagnostics package from the iDRAC web interface. Then power cycle, press F11 and choose System Utilities > Launch Diagnostics. |
The ePSA Pre-boot System Assessment window lists every detected device and starts testing them. Press Esc to stop testing. Click Exit at the top of the screen when you finish. If you restart without exiting, the Lifecycle Controller menu loads again on the next boot.
Step 3 — Run a full or targeted test
The default run tests all detected devices. To focus on one suspect component, as described in the Lifecycle Controller User's Guide:
- Start the diagnostics from Lifecycle Controller as above.
- Pause testing, click the right arrow, then click Advanced Options.
- Deselect the components you do not want to test.
- Set Test Iterations. 0 means it repeats until you stop it.
- Close the dialog, select Thorough Test Mode, and click the centre arrow to start.
Thorough tests add detail about failed devices. They take longer, so use them on the suspect part rather than on the whole server.
Step 4 — Read the results
| Menu | What it shows |
|---|---|
| Configuration | Configuration and status of all detected devices |
| Results | Results of all tests that ran |
| System health | Current overview of system performance |
| Event log | Timestamped log of all test results (shown once at least one event is recorded) |
For each failure, record the error code and message exactly, together with the device and its location (DIMM slot, drive bay, PSU number). Dell publishes a reference table of ePSA and PSA error codes (2000-0NNN series) with troubleshooting steps in its knowledge base. Search for your exact code on Dell support and follow the steps listed for it. If the test confirms a failed part, escalate to Dell support (see below).
Remote diagnostics through iDRAC
When nobody is on site, iDRAC9 can schedule the same offline diagnostics and keep the results. This needs at least an iDRAC Express licence, Lifecycle Controller enabled, and the Login and Server Control privileges. Only one diagnostic job can be scheduled or running at a time.
racadm diagnostics run -m 1 -r forced -s TIME_NOW -e TIME_NA
racadm getremoteservicesstatus
racadm diagnostics export -f diags.txt -l //192.0.2.10/share -u user -p pass-m:0Express (a subset of tests),1Extended (all tests),2both in sequence.-r:pwrcycle,Graceful(waits for the OS to shut down) orForced(asks the OS to shut down and power cycles if it has not after 10 minutes).-sand-etakeyyyymmddhhmmss.TIME_NOWschedules the job immediately andTIME_NAmeans no expiry. With a scheduled window, keep more than five minutes between start and end.- Export results only when the server is out of POST and
getremoteservicesstatusshows LC Ready. The export supports CIFS, NFS, HTTP and HTTPS, up to 5 MB.
The run command returns a job ID, and the diagnostic events and results are recorded in the Lifecycle Log. A Scheduled job can be cancelled. To stop one that is already running, restart the server.
When to escalate to Dell
If ePSA reports a failure, or the server has symptoms but passes, collect a SupportAssist Collection. It includes hardware inventory, firmware versions, logs and storage controller logs in one ZIP. You can export it from Lifecycle Controller (Hardware Diagnostics > Export SupportAssist Collection) or from iDRAC (see SupportAssist collection from iDRAC9). Send it to Dell support together with the ePSA error code and the Service Tag.
Frequently asked questions
Does ePSA error 2000-0251 mean a hardware failure?
No. Dell says it means the event log contains old unchecked records. Save the log, clear it, and run the diagnostics again.
Can I run diagnostics without going to the server?
Yes. Use the iDRAC Virtual Console to press F10 or F11, or schedule remote diagnostics with racadm diagnostics run.
Checklist
- Downtime window agreed and the SEL and Lifecycle Log exported
- ePSA launched from F10 or F11 (or scheduled remotely)
- Targeted thorough test run on the suspect component
- Error codes, devices and slots recorded exactly
- Steps from Dell's reference for each error code followed
- SupportAssist Collection and the Service Tag sent to Dell for confirmed failures
Skyline can supply, configure and maintain Dell, HPE and Huawei servers under an annual maintenance contract (AMC). Contact us.
Comments
0 total · 0 threads