Data corruption after backup EncFS directory

[ShredRum]

Data corruption after backup EncFS directory

Description:

I am experiencing data corruption issues after performing a backup of my EncFS directory. Here are the details of the issue:

• During mounting, I encounter the following error for many files:
  WARNING MAC comparison failure in block 0
  As a result, the files are not readable at all.
• When using the --forcedecode option, some files can be partially read, but I receive numerous errors related to different block numbers during access attempts.

---

Steps to Reproduce:

1. Create an EncFS encrypted folder using the following configuration:
   • Cipher algorithm: ssl/aes
   • Key size: 256 bits
   • Block size: 1024 bytes
   • Filename encoding: nameio/block (with IV chaining enabled)
   • KDF: PBKDF2 with 100,000 iterations
2. Copy the encrypted directory to another location as a backup.
3. Attempt to access the backed-up files or restore them to a new location.

---

Observed Behavior:

• Files in the backed-up directory are severely corrupted.
• Many files appear to have a fixed size of 100 bytes in the mounted view.
• Attempts to open the files result in readable headers, but the rest of the content is unreadable or contains junk data.
• Directory names and structures appear intact.

Troubleshooting Attempts:

1. Verified the original .encfs6.xml is intact and matches the backed-up directory.
2. Attempted recovery using the same password and EncFS version but could not restore the files.

---

Request:

Could you help determine if:

• This is a known issue or configuration error.
• There are any recommended recovery steps for partially corrupted files.

Thank you for your help!

Example of reading a file with the --forcedecode argument

[rfjakob]

Hi, for reproduce step (3) you mean the encrypted view? Otherwise i don't understand.

…

On Fri, 29 Nov 2024, 21:11 ShredRum, ***@***.***> wrote: Data corruption after backup of mounted EncFS directory Description: I am experiencing data corruption issues after performing a backup of my EncFS directory. Here are the details of the issue: - During mounting, I encounter the following error for many files: WARNING MAC comparison failure in block 0 As a result, the files are not readable at all. - When using the --forcedecode option, some files can be partially read, but I receive numerous errors related to different block numbers during access attempts. ------------------------------ Steps to Reproduce: 1. Create an EncFS encrypted folder using the following configuration: - Cipher algorithm: ssl/aes - Key size: 256 bits - Block size: 1024 bytes - Filename encoding: nameio/block (with IV chaining enabled) - KDF: PBKDF2 with 100,000 iterations 2. Mount the EncFS directory. 3. Copy the mounted directory (the decrypted view) to another location as a backup. 4. Attempt to access the backed-up files or restore them to a new location. ------------------------------ Observed Behavior: - Files in the backed-up directory are severely corrupted. - Many files appear to have a fixed size of 100 bytes in the mounted view. - Attempts to open the files result in readable headers, but the rest of the content is unreadable or contains junk data. - Directory names and structures appear intact. Troubleshooting Attempts: 1. Verified the original .encfs6.xml is intact and matches the backed-up directory. 2. Attempted recovery using the same password and EncFS version but could not restore the files. ------------------------------ Request: Could you help determine if: - This is a known issue or configuration error. - There are any recommended recovery steps for partially corrupted files. - Backing up a mounted EncFS directory is not supported or requires special handling. Thank you for your help! Example of reading a file with the --forcedecode argument Screenshot.2024-11-29.220713.png (view on web) <https://github.com/user-attachments/assets/0f96bbe2-b5f9-4e31-9079-ab5a6f8f6d2b> — Reply to this email directly, view it on GitHub <#672>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AACGA77RZEECJDOFGWSG35D2DDC5ZAVCNFSM6AAAAABSXWEVAKVHI2DSMVQWIX3LMV43ASLTON2WKOZSG4YDMMBVGMYDKOA> . You are receiving this because you are subscribed to this thread.Message ID: ***@***.***>

[ShredRum]

Hi, for reproduce step (3) you mean the encrypted view? Otherwise i don't understand.
…
On Fri, 29 Nov 2024, 21:11 ShredRum, @.> wrote: Data corruption after backup of mounted EncFS directory Description: I am experiencing data corruption issues after performing a backup of my EncFS directory. Here are the details of the issue: - During mounting, I encounter the following error for many files: WARNING MAC comparison failure in block 0 As a result, the files are not readable at all. - When using the --forcedecode option, some files can be partially read, but I receive numerous errors related to different block numbers during access attempts. ------------------------------ Steps to Reproduce: 1. Create an EncFS encrypted folder using the following configuration: - Cipher algorithm: ssl/aes - Key size: 256 bits - Block size: 1024 bytes - Filename encoding: nameio/block (with IV chaining enabled) - KDF: PBKDF2 with 100,000 iterations 2. Mount the EncFS directory. 3. Copy the mounted directory (the decrypted view) to another location as a backup. 4. Attempt to access the backed-up files or restore them to a new location. ------------------------------ Observed Behavior: - Files in the backed-up directory are severely corrupted. - Many files appear to have a fixed size of 100 bytes in the mounted view. - Attempts to open the files result in readable headers, but the rest of the content is unreadable or contains junk data. - Directory names and structures appear intact. Troubleshooting Attempts: 1. Verified the original .encfs6.xml is intact and matches the backed-up directory. 2. Attempted recovery using the same password and EncFS version but could not restore the files. ------------------------------ Request: Could you help determine if: - This is a known issue or configuration error. - There are any recommended recovery steps for partially corrupted files. - Backing up a mounted EncFS directory is not supported or requires special handling. Thank you for your help! Example of reading a file with the --forcedecode argument Screenshot.2024-11-29.220713.png (view on web) https://github.com/user-attachments/assets/0f96bbe2-b5f9-4e31-9079-ab5a6f8f6d2b — Reply to this email directly, view it on GitHub <#672>, or unsubscribe https://github.com/notifications/unsubscribe-auth/AACGA77RZEECJDOFGWSG35D2DDC5ZAVCNFSM6AAAAABSXWEVAKVHI2DSMVQWIX3LMV43ASLTON2WKOZSG4YDMMBVGMYDKOA . You are receiving this because you are subscribed to this thread.Message ID: @.>

Hi, yep. When I try to decrypt the backup in step 3, I gotten corrupted files (in decrypted form).

[rfjakob]

Sounds like the copy operation somehow got interrupted or messed up. But encfs is not involved in this step.

The original is fine I guess?