Server memory uses error-correcting code (ECC). When the system detects and corrects an error, it logs a correctable error and keeps running. When the error cannot be corrected, it logs an uncorrectable (multi-bit) error, and the operating system or application may crash. Dell's KB puts it this way: a correctable event "is generated when there is a memory error correction on a particular Dual Inline Memory Module (DIMM)", and an uncorrectable event "is generated when the chipset is unable to correct the memory errors". The two need different urgency, and both vendors document what to do for each.
Before you start
- Record the exact message ID or text and the DIMM location (socket and slot) from the iDRAC Lifecycle Log or the iLO IML.
- Check whether the errors point at one DIMM or several. Dell's KB treats single-DIMM and multiple-DIMM errors differently.
- Note the current BIOS/system ROM and management controller firmware versions.
- Plan a maintenance window: reseating or replacing DIMMs requires powering the server off and removing all power.
Warning: Dell's troubleshooting steps include clearing the System Event Log so new messages stand out. Export or save the SEL (or a SupportAssist collection) before clearing it, because cleared entries cannot be recovered. Uncorrectable errors can crash the OS, so make sure workloads and data are protected before you start; back up first.
Step 1 — Find the message and the DIMM
- Dell: iDRAC9 Maintenance > Lifecycle Log and System > Overview > Memory. On the R760, the memory indicator on the front panel turns solid amber when a memory error occurs; Dell's action is to check the System Event Log for the location of the failed memory and reseat it.
- HPE: iLO Information > Integrated Management Log and System Information > Memory. The Physical Memory list shows a status per DIMM, including Degraded, Configuration Error, Map Out Error (mapped out because of a training failure) and Not Supported. On Gen11, a correctable threshold problem can also show as an amber memory LED.
Step 2 — Match the message to the documented action
| ID and severity | Message and recommended action |
|---|---|
| MEM0701 (Info), MEM0702 (Warning) | "Correctable memory errors detected". Memory is operational. Reboot to enable potential memory self-healing, make sure BIOS is up to date, and collect a SupportAssist collection if it persists. |
| MEM0802 | The memory health monitor detected degradation in a DIMM. Restart the system to start the self-heal process. |
| MEM0804 / MEM0805 | Self-heal completed successfully (no action) / self-heal could not complete: replace the DIMM during the next planned service. |
| MEM0000 (Info), MEM0005 (Critical) | Persistent correctable errors / persistent correctable error limit reached. Described as an early indicator of a possible future uncorrectable error. Reinstall the memory; collect a SupportAssist collection if it persists. |
| MEM0001 (Critical) | "Multi-bit memory errors are detected on the memory device at location(s) ... Immediately replace the DIMM." |
| MEM0007 (Critical) | Unsupported memory configuration; memory size is reduced. Review the supported memory configurations. |
| Message | Recommended action |
|---|---|
| "Correctable Memory Error Threshold exceeded" (class 0x000a, code 0x0460) | Causes include poor contact between DIMM and socket and DRAM degradation. Remove and reinstall the DIMM, update system firmware, and if it persists open a support case and replace the DIMM. |
| "High rate of corrected memory errors, performance may be degraded" (0x000a / 0x0461) | Reseat the indicated DIMM. |
| "DIMM failure - Uncorrectable memory error" (0x000a / 0x0284) | At the next maintenance window power down, remove all power, remove the DIMM, clean the DIMM and slot with anti-static compressed air, reinstall; open a case if it persists. |
HPE's Gen11 troubleshooting guide lists the symptoms of an uncorrectable error: a POST or IML message, a Windows stop error, a purple diagnostic screen, a Linux kernel panic, a hang or freeze, ASR, unexpected restarts or power-downs, and parity errors.
Step 3 — Correctable errors: update, restart, reseat
- Update firmware. Dell's KB lists CPLD, iDRAC and BIOS; HPE's guidance is to update the system ROM (for example with the SPP).
- Dell: if the message asks for it, restart so the BIOS can attempt self-healing, then check for MEM0804 or MEM0805. The R760 BIOS lists System BIOS > Memory Settings > DIMM Self Healing (Post Package Repair) on Uncorrectable Memory Error (Enabled by default), ADDDC Settings (x4 DIMMs) and Correctable Error Logging, which that manual lists as Disabled by default.
- If the errors continue, power the server off, remove all power and reseat the reported DIMM, following your model's service manual and DIMM handling guidelines.
- Dell: after reseating, check that System Setup Main Menu > System BIOS > Memory Settings shows the expected System Memory Size, then run the memory test in system diagnostics (see PowerEdge hardware diagnostics).
- If correctable errors keep returning on the same DIMM after these steps, replace it under warranty.
Step 4 — Uncorrectable errors: reseat, then replace
- Dell: MEM0001 tells you to replace the DIMM immediately. Dell's KB says single-DIMM errors should be checked against the Lifecycle Controller log and its recommended actions, and multiple-DIMM errors should go to Dell support.
- HPE: reseat the DIMM, update the system ROM, and replace the DIMM if the issue still exists.
- If you are unsure which DIMM is at fault, HPE's isolation method is to test each channel by removing the other DIMMs, then swap suspect DIMMs with known-good ones.
Step 5 — Rule out configuration problems
Some memory errors are configuration errors, not failing DIMMs. HPE lists causes such as channels populated in the wrong order, unsupported DIMMs, a missing processor for the populated slots, or an Advanced Memory Protection mode the DIMMs do not support. Dell reports MEM0007 for an unsupported configuration. Check the population rules for your platform: PowerEdge memory population rules or ProLiant memory population rules.
When to escalate
Open a vendor case for any uncorrectable error that returns after reseating, for errors on several DIMMs at once, or when self-heal fails. Attach a SupportAssist collection (Dell) or an Active Health System log (HPE).
Frequently asked questions
Do I need to replace a DIMM after one correctable error?
Not according to the messages above. Dell's correctable messages say the memory is operational and recommend a restart for self-healing and a BIOS update; replacement comes when errors persist, self-heal fails or a threshold is reached.
Can I move the DIMM to another slot to test it?
HPE documents swapping suspect DIMMs with known-good ones to isolate a failure. Keep the population rules in mind so the test does not create a configuration error.
Checklist
- Message ID, severity and DIMM location recorded; SEL exported before any clear.
- Firmware updated (BIOS/system ROM, iDRAC or iLO, CPLD where listed).
- Correctable: restart for self-heal (Dell), reseat if errors continue.
- Uncorrectable: DIMM reseated or replaced; memory size and diagnostics checked.
- Population rules and memory protection mode confirmed.
- Case opened with logs when errors persist or affect several DIMMs.
Skyline can supply, configure and maintain Dell, HPE and Huawei servers under an annual maintenance contract (AMC). Contact us.
Comments
0 total · 0 threads