Intel RST Says Delete Volume. That Wipes the RAID Metadata.

Written and maintained by the Desert Data Recovery lab team. · Updated October 1, 2026

We already wrote about Initialize Disk, about the Rebuild button, about a RAID 1 mirror left degraded until the survivor was the only copy, about a RAID 0 stripe when one drive dies, about Force Online after a hardware RAID battery fails, and about Storage Spaces Delete Pool. This page is the Intel volume. RST or VROC still has the disks, or it says it does not, and the offered fix is Delete Volume or Reset to non-RAID.

What you are looking at

Intel Rapid Storage Technology is motherboard RAID. An option ROM or a UEFI driver, plus a Windows driver, assembles member disks into one volume. The definition of that volume — RAID level, which disks belong, stripe size, member order, and on a mirror which disk is current — is stored on the members. Intel VROC (Virtual RAID on CPU) is a later Intel volume manager, usually in front of NVMe through VMD. It is not the same product, and a board does not offer every RAID level just because the logo says Intel. The delete and reset actions are the same class of write.

The screen is one of these:

• The Windows app. Intel Rapid Storage Technology on older installs. Intel Optane Memory and Storage Management on newer ones. The volume line says Normal, Degraded, Failed, or the volume is simply gone. Member disks are listed under it, or one of them says failed or missing.
• The pre-boot utility. Older boards still interrupt POST for a RAID option ROM (the Ctrl+I screen): Create RAID Volume, Delete RAID Volume, Reset Disks to Non-RAID. Newer boards bury the same choices in UEFI setup, or in a VROC page. The words move. The write does not.
• A volume that disappeared after a BIOS change. SATA mode flipped from RAID to AHCI, VMD was turned off, a BIOS update dropped the RAID driver, or CMOS was cleared. Windows no longer shows one disk. The members can still be carrying the metadata. The forum reply is to reset them to non-RAID so Disk Management can see them. That reply deletes the metadata you would need the mode switch not to have touched.

RST volumes that show up on a bench are RAID 0, RAID 1, RAID 5, and RAID 10, when the board actually offers that level. Older Intel utilities could also put two volumes on the same disks — a small mirror and a stripe on the capacity left over. Deleting "just the broken one" still rewrites metadata on disks the other volume uses. VROC's menu depends on the platform. If the screen names a RAID level and then offers Delete or Reset, the level does not make the click safer.

This is not an Intel Optane H10. That module is Optane plus QLC NAND on one M.2. The Optane app can be installed on a machine that also has an H10, which is how the two problems get mixed up. A RAID volume across separate SATA or NVMe disks is this page. One hybrid stick is that page.

What Delete Volume and Reset to non-RAID do

Delete Volume removes the volume definition from the member disks. The option ROM calls it Delete RAID Volume. The Windows app calls it Delete Volume. Intel's dialog tells you the data on that volume will be lost. People click through it because the volume is already missing and the sentence sounds like a description of the failure. It is a description of the click.

What the click removes is the RAID metadata: that these disks were one volume, which level, how wide the stripe was, which port was first, and — on a mirror that has been degraded — which member received the writes after the other one dropped. The file bytes may still sit on the platters or the NAND after that metadata is gone. The firmware no longer has a volume to hand you. The next screen is a new volume, or Windows offering to initialize the disks one at a time.

Reset to non-RAID is the same metadata write aimed at the disks you select. The option ROM label is Reset Disks to Non-RAID. People use it when the volume will not mount and they want Windows to see the members as ordinary disks. It works, in the narrow sense that the RAID flag comes off. The members can then show up as Unknown, Not Initialized, or unallocated. That is the Initialize Disk dialog, and confirming it writes a new partition table. Reset did not extract the files. It made the disks look empty enough for Windows to offer a wipe of the map.

On a RAID 0 stripe, order and stripe size are the volume. The RAID 0 page already says a reset to non-RAID leaves members that can still be readable and a volume that is not, until that metadata is put back from copies — or until someone initializes the disks and that chance is gone. Delete Volume is how the metadata leaves.

On a RAID 1 mirror, both disks once held the volume, and they stop being copies the day one member falls out. Every later write exists on the survivor only. Reset to non-RAID drops the record of which disk that was. A forum note that "deleting a RAID 1 keeps the data" is about a healthy mirror someone intended to break. A degraded or failed RST mirror is not that job. The RAID 1 page is the weeks the warning sat there. This page is the Delete and Reset buttons on the RST screen, including on that mirror.

On RAID 5 and RAID 10 the files are not whole files on one disk. Parity, or a mirror pair inside a stripe, only means something with the layout the metadata recorded. Deleting the volume does not convert a member into a drive letter full of documents.

What not to do

Do not click Delete Volume to clear a failed, missing, or degraded status. The status is the firmware telling you it cannot present the volume it still has a definition for, or that it is presenting the volume with a member already out. Deleting the definition is not a diagnostic. It does not read the member that failed. It does not copy the share to another disk.

Do not reset disks to non-RAID so Windows, Disk Management, or a USB dock will see them. That is the click this page is about. A member moved to another PC "to check if the good drive is OK" gets the same offer: initialize, or reset, or both. Slide it back into the original port afterward and the set no longer matches what the other disks still say.

Do not create a new RST or VROC volume on these disks. Create RAID Volume writes a new definition. On a stripe that is a new layout. On a mirror the firmware will usually initialize or resync, which picks a direction and writes. Windows setup that offers to create a volume so the install can continue is the same wizard. Empty disks are where a new volume belongs, after a copy you trust already exists somewhere else. These disks are not empty just because the letter disappeared.

Do not initialize or format a member. After a reset, or after AHCI mode made the volume disappear, Disk Management shows Unknown / Not Initialized. Initialize Disk writes a new signature and partition table. On GPT it also writes a backup table at the end of the disk, which is where Intel RAID metadata is stored on the member. Initialize can overwrite both the front of the disk and that metadata. Format, a new simple volume, and diskpart clean continue the job. Cancel the dialog.

Do not rebuild, repair, or resync as a test while you are still deciding whether the volume is healthy. A rebuild reads the members the firmware trusts and writes a replacement. That is the maintenance job on the rebuild page only when one disk has just failed, the volume is online, and the remaining members are confirmed healthy. A failed RST volume, a missing volume, or a mirror that has been degraded for weeks is not that sentence. Delete Volume is not a substitute for figuring out which of those you have. If a rebuild is already running and a disk is clicking or dropping out, power the machine down. Finishing the pass is more writes.

Do not keep flipping SATA mode, VMD, or CMOS to see which setting makes the letter return. Photograph the mode you are in now — RAID or AHCI, VMD on or off — and stop. Toggling is how people end up in the option ROM with Delete Volume highlighted. Clearing CMOS again does not put metadata back once a reset has removed it.

Do not run chkdsk /f or Error Checking because the volume still has a letter and RST says Degraded. Degraded is a member, not a dirty NTFS bit. What /f does is the CHKDSK post. It writes the filesystem on the disk that is still serving.

Why this is wipe-class

A full format and a metadata reset are not the same number of sectors. The honest description is still wipe-class for recovery. The thing a lab reads first on an Intel volume is the RAID metadata that ties the members into one disk. Delete Volume and Reset to non-RAID destroy that record on the disks themselves. Intel's own warning treats the volume's data as deleted. There is no undo in the app.

The file bytes are often still on the media at that moment. That is why a second click matters. Initialize replaces the partition map. A new volume writes a new RAID header and then synchronizes or stripes. A format writes an empty filesystem. Each of those is how "the metadata is gone but the bytes might remain" becomes "the bytes were overwritten." People do the sequence because the first click did not bring the files back and the next dialog looks like the repair.

A degraded mirror has a sharper version of the same problem. The two members already disagree. The metadata is the note about which one is current. After a reset, both disks are just disks, and a resync or a new mirror can copy the stale one onto the survivor. That direction is not a button you get back.

This screen is not the other pages

Optane H10 is one M.2 with two devices on it. The published H10 case is that module. If Disk Management shows a single hybrid stick and there is no RAID volume across separate disks, use that post. If Optane Memory and Storage Management lists member disks under a RAID volume, you are here. Do not follow the H10 page into a clone of "the SSD," and do not follow this page as if an H10 were a two-disk mirror you built.

RAID 0, one drive dies is the stripe: one dead member and no second copy. That post already names RST reset-to-non-RAID as the way a readable pair becomes a metadata problem. Use it for what a stripe is. Use this page when the button in the RST or VROC UI is Delete Volume or Reset, including on a stripe that has not lost a member yet and is simply missing after a BIOS change.

The RAID 1 page is a mirror that kept serving after one member failed, and the warning sat there until the survivor or a rebuild spent the only current copy. An RST mirror can be that story. The metadata and the Delete / Reset labels are this page. If the firmware says RAID 1, the share is still up, and the question is whether to rebuild a disk that has been the whole array for weeks, read that page and do not start the rebuild. If the menu in front of you says Delete Volume or Reset to non-RAID, stop and stay here.

Rebuild vs. recovery is the general difference: a rebuild writes a replacement from the members the controller believes are healthy, and a lab images first. RST will offer a rebuild on a degraded redundant volume. That offer is not Delete Volume. Do not delete the volume because you are unsure about the rebuild, and do not rebuild because Delete sounded worse. If the array is offline, more than one disk is involved, or you cannot say the rest are healthy, neither click belongs on the originals.

RAID 5 with two drives down is a parity set that has already lost a second member. An RST RAID 5 that failed a second disk is that shape of failure on Intel metadata. The "do not rebuild on the originals" rule transfers. The option ROM's Delete Volume does not. That post will not tell you what Reset Disks to Non-RAID changes.

The RAID battery post is a PERC, SmartArray, MegaRAID, or the same class of hardware card after a power cut. Force Online and Clear Foreign Configuration are that screen. Clear Foreign drops controller metadata the way Reset to non-RAID drops Intel metadata. They are not the same utility. If your screen is a battery, preserved cache, or a foreign configuration on a hardware card, use that post. If the logo is Intel RST or VROC, you are here.

Storage Spaces is a Windows pool. Delete Pool, Remove Disk, and Reset are that metadata. The Control Panel or Server Manager wording is the tell. RST does not call the set a storage pool. If both are installed, read which window is actually open before you follow either article. Initialize Disk is the dialog after a member has already been reduced to an ordinary disk. Do not reset the array in order to reach it.

A Synology, QNAP, Unraid, or TrueNAS box is not this firmware. Volume Crashed, Volume Not Active, Unmountable, and a pool that will not import are Synology, QNAP, and Unraid / TrueNAS. Do not hunt those menus for a Reset to non-RAID click, and do not run Delete Volume on an Intel board because a NAS article said the volume was crashed.

What to do right now

Stop clicking Delete Volume. If a disk is clicking or the rebuild is already grinding, power the machine down. If the disks are quiet and the screen is still up, photograph it, then power down. Do not initialize the members. Do not create a new volume on them. Do not reset them to non-RAID to make Windows see them. The set comes in as it is.

1. Stop. Cancel Delete Volume, Reset to non-RAID, Create Volume, and Initialize. If a disk is clicking, grinding, or dropping out, or a rebuild is running and the source is struggling, power the machine down. If everything is quiet, stop changing BIOS mode and do not start a rebuild as a test.
2. Photograph the screen before the next reboot changes the words. App or option ROM. RAID 0, 1, 5, or 10. Normal, Degraded, or Failed. Which port, slot, or serial is missing. The exact button offered. SATA mode or VMD, if you are in BIOS setup. Intel RST, Optane Memory and Storage Management, or VROC — whichever title is in the corner. Screenshots help. They are not a diagnosis.
3. Label every member before anything is unplugged. SATA port, M.2 slot, or bay, and the serial if you can read it without pulling the drive. Keep the disk RST already marked failed or non-RAID, the ones it still calls members, and any spare you inserted. Member order is part of the volume. A pile of unlabeled sticks is how the wrong one gets initialized on the bench later.
4. Write down what you remember: whether anyone already deleted a volume, reset disks to non-RAID, initialized, formatted, created a new volume, or started a rebuild. Whether SATA mode or VMD was changed, or CMOS was cleared. And whether a backup exists on media that is not this set. "The RAID is the backup" means it does not. If the volume was BitLocker or device encryption, the key still has to exist later. Write down that you have it.
5. Leave the originals alone. Do not save files back onto a member. Do not move a disk to a USB dock or another motherboard so Disk Management can look at it. Copying off is reasonable only while the volume is still mounted, the disks stay quiet, and the copy is going to different media. If the volume is failed or missing, a disk is clicking, or a reset already happened, that is an evaluation, not a copy.

Pack loose disks the way you would any failed drive — how to ship a failed drive safely. If a member is clicking, start with power it off. A desktop whose M.2 drives you cannot pull cleanly can come in as a chassis. Do not open a drive on a table to see which one the option ROM called failed.

How a lab handles it

We do not delete the volume, reset the disks to non-RAID, or create a new RST or VROC volume on your hardware to see if a letter comes back. We do not initialize a member in Disk Management, and we do not rebuild the set on the original controller as a test.

Each member is imaged write-blocked first. That includes a disk the UI already calls failed, missing, or non-RAID, and any spare a rebuild had started writing. A drive that only reads in pieces is still imaged as far as it will go. Imaging uses the lab tools already named on this site, including the ACE Lab PC-3000 and DeepSpar Disk Imager. Work happens on the copies.

On those images the question is what the remaining metadata still says, and what the filesystem on each member says when the metadata has already been stripped. Level, order, stripe size, and which mirror member is current are answered from the copies. A wrong guess costs a clone. The same guess in the option ROM writes the disks. Where a reset has removed the RAID header, reconstruction uses what is still on the media. It does not put the header back by clicking Create Volume on the originals.

What imaging cannot do: put back a RAID map a new volume already replaced, a member that was initialized and then formatted, or a resync that copied the stale mirror side onto the current one. We will tell you if that current copy is not there.

The public case log has no Intel RST volume, no VROC volume, and no case where Delete Volume or Reset to non-RAID was the click. We will not invent one.

The only Intel device in that log is a different object. DDR-2026-0009: an Optane Memory H10, Optane plus QLC NAND on one M.2. The QLC NAND side was not recoverable. The Optane portion was imaged, decrypted, and the file system rebuilt from that image. A large majority of the files the client needed came back. Partial recovery. It is not an RST RAID set, and it is not a Delete Volume intake.

The closest published array work is other firmware. DDR-2026-0003: a Synology four-drive RAID 5, two members failed, nothing rebuilt in the NAS, each drive imaged, the array reconstructed from the images. Full recovery. 21607: twelve HPE 1.8 TB SAS drives on one SmartArray, a five-drive RAID 5 and a seven-drive RAID 6, three RAID 6 members failed, nothing rebuilt on the original controller, every member imaged, the arrays reconstructed virtually. Full recovery. DDR-2026-0014: a two-bay NVMe enclosure, two 4 TB WD_BLACK SN850X SSDs as one striped APFS volume, both members failing at the NAND, imaged first, the stripe worked out from the copies. Full recovery of that client's data. The enclosure is not Intel RST. Those outcomes belong to those drives. They are not a rate, and they are not a prediction about a volume whose RST screen is offering Delete. We don't advertise a success rate.

If the stick is an H10, start with two drives on one module. If the firmware says stripe and one member is already dead, start with RAID 0. If the button is Rebuild and you are not sure the rest of the set is healthy, start with rebuild vs. recovery. For the work itself, see RAID data recovery.