[BUG] Filesystem corruption: "ext3_new_block: Allocating block in system zone"

Filesystem corruption: "ext3_new_block: Allocating block in system zone"

Issue:

• Filesystem corrupted with following messages in the log (note, message format modified for this newsletter):

  kernel: EXT3-fs error (device dm-22): ext3_new_block:
  Allocating block in system zone - blocks from 4 816900, length 1
  kernel: Aborting journal on device dm-22.
  kernel: ext3_abort called.
  kernel: EXT3-fs error (device dm-22): ext3_journal_start_sb:
  Detected aborted journal
  kernel: Remounting filesystem read-only
  kernel: EXT3-fs error (device dm-22) in ext3_reserve_inode_write:
  Journal has aborted
  kernel: __journal_remove_journal_head: freeing b_committed_data
  kernel: __journal_remove_journal_head: freeing b_committed_data
  kernel: ext3_abort called.
  kernel: EXT3-fs error (device dm-22): ext3_remount:
  Abort forced by user
  kernel: ext3_abort called.
  kernel: EXT3-fs error (device dm-22): ext3_remount:
  Abort forced by user
  kernel: ext3_abort called.
  kernel: EXT3-fs error (device dm-22): ext3_remount:
  Abort forced by user
             

Environment:

• Red Hat Enterprise Linux (RHEL) 5
• Red Hat Enterprise Linux (RHEL) 6

Resolution:

• Tracked by internal Red Hat Bugzilla 538943
• NOTE: Be sure to unmount the filesystem before doing this repair step:
• Use fsck on the file system to fix the corruption. The following command can be used:

 #e2fsck -fyv 
              

• The fix for this issue is currently proposed for future versions of RHEL

Root Cause:

The filesystem block allocator tried to allocate a system zone block (i.e. a metadata block) with user data. This happens when the bitmap of that specific system zone block is corrupted (usually set as free).

Back to top...