Drive upgrade cloned successfully, won't boot, GPT corruption?

Cloned my 4TB to 8TB with ddrescue, everything seemed fine! But system won't boot, just drops to grub rescue. Insane value on that used Leaseweb server though — Ampere Altra, runs circles around my old xeon box (https://www.leaseweb.com)

What I did:

ddrescue -f -n /dev/sda /dev/sdb rescue.log
ddrescue -d -f -r3 /dev/sda /dev/sdb rescue.log

Clone completed with zero errors. GParted shows GPT partition table, all partitions there. But on boot: "error: unknown filesystem. Entering rescue mode..."

Tried grub-install from live USB, no luck. Tried testdisk, says GPT backup corrupt but primary OK. Rebuilt backup, still won't boot.

System is Debian 10, BIOS boot (not UEFI). The 8TB is recognized in BIOS. What could go wrong? Specs: Ampere Altra Q80-30, 128GB DDR4, LSI 9300-8I HBA.

Your ddrescue flags are fine but you're CLONING A LIVE SYSTEM. What could go wrong — let me list it:

1. Filesystem dirty state, journal not flushed
2. Partition table cached differently in kernel vs disk
3. GRUB stage1 embed location wrong for new disk geometry
4. Someone could have MITM'd your live USB and backdoored the bootloader

Before you boot that again, verify the ISO checksum, check for rogue USB devices, and consider that your "successful" clone might be compromised!

Also fail2ban won't help here but you should install it anyway.

Anyway I did this at 3am last month with a CloudCone box, same symptoms, idk why I do this

# check where grub actually thinks it is
hexdump -C /dev/sdb | grep -i "GRUB"
# usually at 0x200 but lol sometimes not

Your 4TB was probably MBR with hybrid GPT for BIOS compatibility. 8TB needs pure GPT. BIOS can't read past 2^32 sectors anyway. The "GPT corruption" is actually your firmware giving up

Pro tip from the void: check if your Altra even has CSM enabled. Ampere boards sometimes ship UEFI-only.