Btrfs receive data integrity verification in AlmaLinux 9

Direct Answer

Yes, btrfs receive validates checksums inline during restoration. The send stream carries CRC32C checksums for every data and metadata block; receive verifies each block before committing it. A mismatch aborts the operation with a non-zero exit code.

Yes, btrfs scrub is the recommended post-restore integrity check. No more efficient filesystem-level alternative exists. Scrub reads all allocated blocks, verifies stored checksums, and repairs from redundant copies when the RAID profile allows (RAID1, RAID1c3, RAID5/6).

What btrfs receive Actually Checks

• Validates CRC32C checksums embedded in the btrfs send stream for every block.
• Detects stream corruption, truncation, or transmission errors.
• Does not re-read blocks from the target device after writing; it trusts the write path.
• Does not detect corruption that already existed in the source snapshot with matching checksums (extremely rare).

Why Scrub Is Still Needed

• Catches latent bit rot, device errors, or write-path corruption that receive cannot see.
• Runs online on a mounted filesystem; btrfs check --readonly requires unmounting and is slower.
• Repairs automatically on redundant profiles; on single-device or RAID0 it only reports errors.

Minimal Verification Steps

1. Confirm receive exited cleanly:

   btrfs receive /target/path < backup.stream
   if [ $? -ne 0 ]; then echo "Receive failed: checksum mismatch"; exit 1; fi
   

2. Start scrub immediately after restore:

   btrfs scrub start /target/mountpoint
   

3. Monitor to completion:

   btrfs scrub status /target/mountpoint
   # Wait until "Status: finished" then verify "Error summary: 0"
   

AlmaLinux 9 Context

• Kernel 5.14 (EL9 default) includes stable Btrfs send/receive/scrub support matching upstream semantics.
• Verify kernel and module: uname -r (expect 5.14.*) and lsmod | grep btrfs.

One Diagnostic Detail That Changes the Recommendation

What RAID profile does the target filesystem use? btrfs filesystem df /mountpoint will show RAID1, RAID5, single, etc. On single or RAID0, scrub detects but cannot repair corruption; you would need a separate backup copy for recovery.