bbf - a safer and more featureful tool for dealing with bad blocks on harddrives
bbf [options] <instruction> <path>
bbf is a tool built around the workflow in dealing with harddrive bad blocks. It has a number of features to limit risk in using the tool and provides features to more easily track down what files are affected by the bad blocks found. Also gives you the ability to manually mark blocks as corrupted in cases where the block isn't technically bad but causing issues.
- readonly scanning of bad blocks
- safe 'fix' mode which won't overwrite good blocks
- burnin mode for checking new drives
- manual marking blocks as corrupted
- find files given list of blocks
- dump list of files and associated block ranges
- dump list of blocks used by a file
- issue secure drive erasure
- filesystem stressing
- -f, --force : override checking if drive is in use when trying to perform destructive actions
- -t, --rwtype <os|ata> : select between OS or ATA reads and writes (default: os)
- -q, --quiet : redirects stdout to /dev/null or otherwise limits output
- -s, --start-block : block to start from (default: 0)
- -e, --end-block : block to stop at (default: last block)
- -S, --stepping : number of logical blocks to read at a time (default: physical / logical)
- -o, --output : file to write bad block list to (default: $HOME/badblocks.)
- -i, --input : file to read bad block list from (default: $HOME/badblocks.)
- -r, --retries : number of retries on certain reads & writes
- -c, --captcha : needed when performing destructive operations
- -M, --maxerrors : max r/w errors before exiting (default: 1024)
<path> is a block device. Prints out details about the block device.
<path> is a block device. Prints out captcha needed for certain instructions.
<path> is a block device. A read-only scan of the block device for bad blocks. rwtype=ata will be slower but may catch more.
Relevant options: rwtype, start block, end block, stepping, max errors, input file, output file.
<path> is a block device. Writes to bad blocks in an attempt to force the drive to reallocate the block. Attempts to read the block first and will write the read data if successful otherwise it will write zeros. This means it is pretty safe to use even if the blocks 'fixed' aren't in fact damaged.
rwtype=ata will work better.
Requires captcha.
Relevant options: captcha, rwtype, force, input file.
<path> is a file. Gets the list of blocks that a file uses and then goes through each block reading what is there and then writing it back which will force reallocation if a block is bad.
rwtype=ata will work better.
Requires captcha.
Relevant options: captcha, rwtype, retries.
<path> is a block device. Iterates through the blocks of the device performing the following:
- Read block data (zero out on failure)
- Write 0x00's and read back to confirm data integrity.
- Write 0x55's and read back to confirm data integrity.
- Write 0xAA's and read back to confirm data integrity.
- Write 0xFF's and read back to confirm data integrity.
- Write back originally read data.
Requires captcha.
Relevant options: rwtype, start block, end block, stepping, max errors, retries, input file, output file.
<path> is a directory. Spawns a number of threads to hammer the filesystem using a number of functions to stress the filesystem and underlying device. Functions include: create, open, mkdir, unlink, rmdir, write, read, close, readdir, stat, chmod, chown, link, symlink. Cleans up after itself on exit but does consume storage and inodes as it runs.
Use --quiet to keep it from printing out what it is doing and improve performance.
<path> is a non-existant file. Creates a file, expands it to fill up the rest of the filesystem, and spawns a thread per core which writes 1MB blocks to the file at random offsets to stress the filesystem and unerlying device.
<path> is a filesystem mount point. Attempts to find the files associated with any blocks listed in the bad block input file. Useful after running scan to find the files with bad blocks.
Relevant options: input file.
<path> is a filesystem. Scans the filesystem and dumps a list of the files with the blocks on the device it occupies.
<path> is an existing file. Prints out a list of all logical blocks the file uses.
- write-pseudo-uncorrectable-wl
- write-pseudo-uncorrectable-wol
- write-flagged-uncorrectable-wl
- write-flagged-uncorrectable-wol
<path> is a block device. Marks blocks listed in the bad block input file as 'pseudo' or 'flagged' uncorrectable. Blocks marked 'pseudo', when read, cause the drive to perform normal error recovery and return errors if necessary. Blocks marked 'flagged', when read, will simply return errors indicating it is bad. 'wl' means 'with logging' and if read will result in failed reads being stored in SMART logs. 'wol' means 'without logging' and will not log any read failures in the SMART log.
Relevant options: input file.
<path> is a block device. Issues an ATA Security Erase command to the device. What this means specifically is device specific but generally it is supposed to be like a low-level format. Use with care.
Requires captcha.
Relevant options: captcha.
Theoretically a more thorough version of the standard ATA Security Erase command. Similiarly its function depends on the device and may be the same as the regular security erase.
Requires captcha.
Relevant options: captcha.
OS mode is the default but ATA is the suggested mode. Especially for fix and burnin. Use a higher stepping value to improve the performance. The max value depends on the drive but for a 512 logical block size a value of 128 or 256 seems to work well.
When running a fix or burnin, rather than writing zeros like other tools, it will first read the block and try to write it back. This will be non-destructive so long as the same location is not being used at the same time. Only if the block read fails will zeros be used.
A captcha is required for destructive operations. This helps with preventing the accidental running of the tool on the wrong drive.
# bbf info /dev/sdb
/dev/sdi:
- serial_number: XXXXXXXX
- firmware_revision: SC61
- model_number: ST8000VN0022-2EL112
- RPM: 7200
- features:
- form_factor: 3.5"
- write_uncorrectable: 1
- smart_supported: 1
- smart_enabled: 1
- security_supported: 1
- security_enabled: 0
- security_locked: 0
- security_frozen: 0
- security_count_expired: 0
- security_enhanced_erase_supported: 1
- security_normal_erase_time: 698
- security_enhanced_erase_time: 698
- block_erase: 0
- overwrite: 1
- crypto_scramble: 0
- sanitize: 1
- supports_sata_gen1: 1
- supports_sata_gen2: 1
- supports_sata_gen3: 1
- trim_supported: 0
- block_size:
- physical: 4096
- logical: 512
- stepping: 8
- block_count:
- physical: 1953506646
- logical: 15628053168
- size:
- bytes: 8001563222016
- human:
- base2: 7.28TB
- base10: 8.00TiB
# bbf -S 256 -t ata scan /dev/sdb
start block: 0
end block: 15628053168
stepping: 256
logical block size: 512
physical block size: 4096
read size: 131072
Scanning: 0 - 15628053168
Current: 2425512192 (15.52%); bps: 179384.74; eta: 20:26:39; bad: 0
# bbf captcha /dev/sdb
Z8400VR0
# bbf -i ~/badblocks.Z8400VR0 -c Z8400VR0 fix /dev/sdb
# bbf -q fsthrash /mnt/mydrive0
CTRL-C to exit...
^CCleaning up...
# bbf filethrash /mnt/mydrive0/test
Creating file: /mnt/mydrive0/test
Expanding file to fill drive: 200209731584 bytes
Spawning thrashing threads: 4 (one per core)
CTRL-C to exit...
$ git clone https://github.com/trapexit/bbf
$ cd bbf
$ make
...
$ sudo cp -av bbf /usr/local/bin
- github.com: https://github.com/trapexit/bbf/issues
- email: trapexit@spawn.link
- twitter: https://twitter.com/_trapexit
- reddit: https://www.reddit.com/user/trapexit
- discord: https://discord.gg/MpAr69V
This software is free to use and released under a very liberal license. That said if you like this software and would like to support its development donations are welcome.
- Patreon: https://www.patreon.com/trapexit
- PayPal: trapexit@spawn.link
- Bitcoin (BTC): 12CdMhEPQVmjz3SSynkAEuD5q9JmhTDCZA
- Bitcoin Cash (BCH): 1AjPqZZhu7GVEs6JFPjHmtsvmDL4euzMzp
- Ethereum (ETH): 0x09A166B11fCC127324C7fc5f1B572255b3046E94
- Litecoin (LTC): LXAsq6yc6zYU3EbcqyWtHBrH1Ypx4GjUjm