Hot spares explained: global vs dedicated spares on Dell PERC and HPE controllers

A hot spare is an installed, unused drive that the RAID controller can pull into a degraded array automatically, so the rebuild starts without waiting for someone to swap a disk. Dell and HPE both support spares, but they use the same words for different things. This guide explains the spare types on Dell PERC and HPE Smart Array and MR controllers, when each one activates, what happens after you replace the failed drive, and how to choose a spare that the controller will accept.

What a hot spare does

Dell's PERC 11 guide explains that if you have configured hot spares, the controller automatically tries to use them to rebuild a degraded virtual disk. If no hot spare with enough capacity is available, a manual rebuild is needed after you insert a suitable drive. HPE describes the same idea for Smart Array: assigning online spares to an array lets you postpone the replacement of faulty drives. Without a spare, HPE notes, rebuilding begins only when the failed drive is replaced.

Same words, different meanings

Watch the HPE Smart Array wording: its "dedicated" spare can serve several arrays, which is closer to what Dell calls global. Support also varies by model. The PERC 12 H365 series supports global hot spares only (up to four), while the H965 series supports both types.

When the spare activates: failure or predictive failure

• HPE Smart Array SR has a controller-wide spare activation mode. Failure mode (the default) starts the rebuild when a member drive fails. Predictive mode activates the spare when a member reports a predictive (SMART) failure and copies the data while the volume is still healthy. HPE states this is up to four times faster than a typical rebuild, can recover bad blocks and supports all RAID levels, including RAID 0.
• Dell PERC offers Auto Replace Member on Predictive Failure. The controller copies the data from the drive with the SMART error to a hot spare and marks that drive failed only after the copy completes, so the virtual disk does not go degraded.
• HPE MR Gen11 has a Replace Drive on SMART Error setting, and its global spare also activates for a RAID 0 volume when a member reports a predictive failure.

ssacli ctrl slot=1 modify spareactivationmode=predictive
ssacli ctrl slot=1 modify spareactivationmode=failure

These SSA CLI examples come from HPE's guide; replace the slot number with your controller's. Run them as single commands with ssacli, or without the prefix inside the SSA CLI console.

After you replace the failed drive

When the rebuild onto the spare has finished, the spare is now a working member. What happens when you insert a new drive depends on the setting:

• Dell revertible hot spare (Replace Member). If the failed disk is replaced in the same slot, the controller copies the data from the commissioned hot spare to the new disk, and the hot spare returns to standby. If the new disk goes into a different slot, you can run a manual Replace Member. Dell notes a temporary performance impact during the copy.
• Dell persistent hot spare slots (disabled by default). A drive inserted into a hot-spare slot automatically becomes a hot spare with the same properties, unless its protocol or technology does not match.
• HPE dedicated spare. After you install the replacement, the controller restores the data from the activated spare to the new drive.
• HPE auto-replace spare. The spare stays in the array permanently, and the new drive becomes the spare.
• HPE MR. The Persistent Spare and Replace Drive options control the same two behaviours.

Choosing a spare the controller will accept

• HPE Smart Array: the spare must be unassigned or already a spare for another array, the same type as the array's drives (SATA or SAS), and at least as large as the smallest drive in the array.
• HPE MR: a global spare must be the same drive type and at least as large as the failed drive.
• Dell PERC 12: Dell's replacement rules ask for the same protocol and media type, equal or larger capacity, matching encryption (SED or non-SED), equal or higher speed and, for SSDs, the same or a higher endurance class (read-intensive, mixed-use, write-intensive).

If a dedicated hot spare fails during a rebuild, Dell says the rebuild moves to a global hot spare with enough capacity if one exists. Keeping one global spare that matches your largest drives therefore covers more situations.

Where to configure spares

• Dell PERC 11 (HII): Main Menu > Physical Disk Management, select the disk, then Operations > Assign Global Hot Spare or Assign Dedicated Hot Spare, click Go and confirm. Spare behaviour is under Main Menu > Controller Management > Advanced Controller Properties > Spare. Step-by-step: replace a failed disk on Dell PERC.
• HPE Smart Array (UEFI): System Configuration > controller > Array Configuration > Manage Arrays > array > Manage Spare Drives, then Assign Dedicated Spare or Assign Auto Replace Spare. Commands and details: HPE Smart Array spare drives.
• HPE MR Gen11 (UEFI): System Configuration > HPE MRXXX Gen11 > Main Menu > Controller Management > Advanced Controller Properties > Spare.

ssacli ctrl slot=3 array B add spares=1:6
ssacli ctrl slot=1 array A add spares=1e:1:5 sparetype=autoreplace

Frequently asked questions

Hot spare or RAID 6?

Dell's PERC guide states that RAID 6 protects against a second failure while a disk is rebuilding, and that if you use only one array, deploying RAID 6 is more effective than deploying a hot spare disk. With several arrays on one controller, a shared spare protects all of them.

Are hot spares checked while idle?

Yes, on PERC 11 patrol read runs on all configured disks, including hot spares. On HPE MR Gen11, patrol read can run for all spare drives.

Checklist

• Learn your vendor's terms: Dell global and dedicated, HPE SR dedicated and auto-replace, HPE MR dedicated and global.
• Match interface, media type, encryption and capacity to the array.
• Decide between failure and predictive activation (HPE) or enable Auto Replace Member on Predictive Failure (Dell).
• Decide whether you want copyback (revertible or dedicated spare) or a permanent replacement (auto-replace).
• For a single array, compare a spare with RAID 6.
• After any rebuild, put a new spare back in place.

Skyline can supply, configure and maintain Dell, HPE and Huawei servers under an annual maintenance contract (AMC). Contact us.