Working with enterprise-grade WD RE4 drives that previously participated in hardware RAID arrays can present unique challenges when trying to repurpose them. Traditional methods like dmraid or zeroing out drives often fail to properly remove stubborn RAID metadata, especially when encountering CRC errors like 5CD0C0DB.

The specific error message:

ddf1: physical drives with CRC 5CD0C0DB, expected FFFFFFFF on /dev/sda
ERROR: ddf1: Cannot find physical drive description on /dev/sda!

indicates the RAID controller wrote non-standard metadata that doesn't conform to expected patterns. This is particularly common with certain hardware RAID implementations.

When standard approaches fail, try these escalating methods:

Method 1: Enhanced dd Approach

Instead of just zeroing the beginning, overwrite critical areas:

# Wipe first and last 2MB (covers most metadata locations)
dd if=/dev/zero of=/dev/sdX bs=1M count=2
dd if=/dev/zero of=/dev/sdX bs=1M seek=$(( $(blockdev --getsize64 /dev/sdX) / 1024 / 1024 - 2 )) count=2

Method 2: sg3_utils Low-Level Format

WD RE4 drives respond well to SCSI format commands:

sg_format --format --size=512 /dev/sdX
# Wait for completion, then:
sg_format --verify --size=512 /dev/sdX

Method 3: Nuclear Option with hdparm

For particularly stubborn cases:

hdparm --user-master u --security-set-pass pass /dev/sdX
hdparm --user-master u --security-erase pass /dev/sdX

Note: This requires drive support and may take hours for large drives.

After any method, verify with:

mdadm --examine /dev/sdX
dmraid -r /dev/sdX
xxd /dev/sdX | head -n 50

When working with hardware RAID, always:

• Properly disband arrays through controller BIOS first
• Use controller-specific utilities when available
• Document the RAID metadata format used

---

When dealing with RAID-configured drives, especially from hardware RAID controllers or fakeraid setups, the metadata can sometimes become incredibly persistent. Normally, these commands would suffice:

dmraid -r -E /dev/sdX
# Or the nuclear option:
dd if=/dev/zero of=/dev/sdX bs=1M

But with WD RE4 drives, we're seeing a particularly stubborn case where:

• dmraid fails with CRC errors
• Zeroing the drive doesn't fully clear metadata
• Even mdadm formatting in rescue mode doesn't help

The specific error we're encountering suggests deep-level metadata corruption:

ddf1: physical drives with CRC 5CD0C0DB, expected FFFFFFFF on /dev/sda
ERROR: ddf1: Cannot find physical drive description on /dev/sda!
ERROR ddf1: setting up RAID device /dev/sda
no raid disks and with names: "/dev/sda"

After testing multiple methods, here's what actually works for WD RE4 drives:

Method 1: Low-Level Wipe with Special Parameters

# First, identify the exact size of your drive:
blockdev --getsize64 /dev/sdX

# Then wipe with optimized block size:
dd if=/dev/zero of=/dev/sdX bs=1M count=1000
dd if=/dev/zero of=/dev/sdX bs=1M seek=$((($(blockdev --getsz /dev/sdX)/2048)-1000))

Method 2: Using sg3_utils for SCSI Low-Level Format

# Install sg3_utils if needed:
sudo apt-get install sg3-utils

# Perform format:
sudo sg_format --format --size=512 /dev/sdX

Method 3: WD-specific Secure Erase

# Use hdparm for ATA secure erase:
sudo hdparm --user-master u --security-set-pass Eins /dev/sdX
sudo hdparm --user-master u --security-erase Eins /dev/sdX

After any of these methods, verify with:

sudo wipefs -a /dev/sdX
sudo parted /dev/sdX print

For future RAID setups with WD RE4 drives, consider:

• Using --metadata=0.90 with mdadm for better compatibility
• Creating a GPT partition table after RAID removal
• Documenting the exact RAID controller settings used