When a drive in a fault-tolerant array on an HPE Smart Array (SR) controller fails, the logical drive keeps running in a degraded state until a spare or a replacement drive is rebuilt. This guide follows HPE's documented procedure: confirm which drive failed, read the drive LEDs, replace it safely, and watch the automatic rebuild until the array is healthy again.
Before you start
- Make sure the array has a current, valid backup.
- Check the other drives in the same array for errors. HPE's guide says to confirm that no other drive in the array has errors before replacing one.
- The replacement must be the same type as the failed drive (SAS or SATA, hard drive or SSD) and at least as large as the smallest drive in the array. The controller immediately fails a drive that is too small.
- With external storage, the server is the first unit to power down and the last to power up, so drives are not wrongly marked as failed.
Step 1 — Confirm which drive failed
HPE lists these signs of a failed drive:
- The drive status LED is amber.
- If the drive LEDs are not visible, the Health LED on the front of the server lights (it also lights for fan, power supply and temperature problems).
- A POST message lists failed drives at restart, and SSA shows failed drives with a distinctive icon.
- The event appears in the Integrated Management Log (IML) and, with HPE agents, in the Windows event log or Linux system log.
In iLO, open System Information and click the Storage tab to see controllers, volumes and drives. iLO notes that a full set of storage data needs the Agentless Management Service (AMS) running in the OS. From the OS you can also run:
ssacli ctrl all show config
ssacli ctrl slot=0 pd all show statusRecord the bay ID of the failed drive (port:box:bay) and the logical drives that use it. See the IML guide for reading the log entries.
Step 2 — Read the drive LEDs
LED meanings depend on the drive carrier. Check which carrier your server uses in its user guide. On the HPE SmartDrive carrier (documented in the SR Gen10 controller guide):
| LED | State | Meaning |
|---|---|---|
| Locate (blue) | Solid blue | The drive is being identified by a host application. |
| Do not remove (white) | Solid white | Do not remove the drive. Removing it causes one or more logical drives to fail. |
| Do not remove (white) | Off | Removing the drive does not cause a logical drive to fail. |
| Drive status | Flashing green | Rebuilding, migrating, expanding, extending, erasing or spare activation. |
| Drive status | Flashing amber/green | Member of a logical drive and predicted to fail. |
| Drive status | Solid amber | Failed, unsupported or invalid. |
On the HPE Basic drive carrier (documented in the DL380 Gen11 user guide and the SR Gen11 controller guide):
| LED | State | Meaning |
|---|---|---|
| Fault/Locate | Solid amber | The drive has failed, is unsupported or is invalid. |
| Fault/Locate | Flashing amber (1 per second) | Predictive failure alert. Replace the drive as soon as possible. |
| Fault/Locate | Solid blue | Operating normally and being identified by a management application. |
| Online/Activity | Flashing green (1 per second) | Rebuilding, RAID or strip migration, expansion, extension, erasing or spare activation. |
| Online/Activity | Solid green | Online, no activity. |
The Basic carrier table has no "Do not remove" LED, so confirm in SSA or ssacli that the logical drive can tolerate the removal before you pull anything.
Step 3 — Light the locate LED on the right bay
Do not rely on bay labels alone. Turn on the identification LED and check it on the front of the server:
- SSA: select the controller and use Manage Device Identification LEDs (Intelligent Provisioning), or the Identify Device task on the drive.
- UEFI System Utilities: System Configuration > (controller) > Array Configuration > Manage Arrays > (array) > Identify Device, set a duration and click Start.
- ssacli:
ssacli ctrl slot=0 array A modify led=on, thenled=offwhen done.
Step 4 — Replace the drive
- Remove the failed drive from the identified bay.
- Insert the replacement drive in the same bay.
- With the server powered on, all drive activity in the array pauses for 1 or 2 seconds while the new drive initializes. Data recovery (rebuild) then starts automatically.
If you replace a drive while the server is off, HPE's drive procedures say a POST message appears at the next power-up and asks you to press F1 to start automatic data recovery. If you do not, the logical drive stays in a ready-to-recover state and the message repeats at each restart.
You can replace more than one drive at once only within the fault tolerance of the RAID level. For example, in RAID 6 any two drives; in RAID 1 or RAID 10, drives that are not mirrored to each other; in RAID 50, drives in different parity groups. To replace more drives than that, wait until the rebuild completes, as shown by the drive LEDs, before removing the next one.
Warning: Never pull a drive whose "Do not remove" LED is lit, and never remove more drives than the RAID level can lose. If fault tolerance is already compromised (the logical drive has failed), inserting replacement drives does not fix it; HPE advises trying to recover and copy the data first, then replacing drives, recreating partitions and restoring from backup. A drive the controller failed may look healthy after a power cycle or reseat; HPE warns that continuing to use such a marginal drive can lead to data loss.
Step 5 — Monitor the rebuild
- The replacement drive LED flashes green during the rebuild and changes to solid green when it finishes.
- In SSA, the logical drive returns to OK. With ssacli, re-run
ssacli ctrl slot=0 ld all showuntil the status is OK. - The rebuild priority controls how much time the controller gives the rebuild versus normal I/O (Low, Medium, Medium High, High). Change it in SSA under the controller settings or with
ssacli ctrl slot=0 modify rp=high.
The SR controller guide says a rebuild generally needs about 15 to 30 seconds per gigabyte for RAID 5 or RAID 6, depending on I/O load, number of drives, rebuild priority and drive performance. During that time the array is exposed to a second failure; see RAID rebuild time and risks.
If the rebuild stops
If the replacement drive's activity LED goes dark while other drives in the array are active, or its status LED turns amber, the rebuild has ended abnormally. HPE lists three cases:
| What you see | Cause | Action |
|---|---|---|
| No drive in the array has an amber LED | Uncorrectable read error on another drive | Back up what you can, do not remove the drive with the media error, restore data, then reseat the replacement to restart the rebuild |
| The replacement drive is amber | The replacement drive failed | Check capacity and model support, then use a different drive |
| Another array drive is amber | A second drive failed | Follow HPE's recovery steps and restore from backup if needed |
Frequently asked questions
Does the rebuild start by itself?
Yes. With an online spare, it starts automatically when the drive fails. Without a spare, it starts automatically when you insert the replacement drive. See spare drives on Smart Array.
Can I use a larger replacement drive?
Yes. HPE requires a capacity equal to or larger than the smallest drive in the array. The extra space is not used by the existing array.
Checklist
- Backup confirmed and other array drives checked for errors.
- Failed bay confirmed in SSA, ssacli or iLO and with the locate LED.
- Replacement drive is the same type and large enough.
- No drive with "Do not remove" lit was touched.
- Rebuild finished: LED solid green, logical drive OK.
- IML reviewed and the event closed.
Comments
0 total · 0 threads