On Mon, Aug 18, 2003 at 02:13:54PM -0400, Theodore Ts'o wrote: > Given that this sort of failure has been reported at least 2 or 3 > times, now, it's clear we need to address this vulnerability, probably > by keeping a backup copy of the journal inode (or at least the journal > data blocks) in the superblock, so it can survive this particular > lossage mode. This has been implemented, and it's in the development sources of e2fsprogs. Folks who are using debian unstable can get this by installing version 1.34+1.35-WIP-2003.08.21-2 of e2fsprogs. Folks who are willing to compile e2fsprogs from source can grab the 1.35-WIP-2003-08-21 Work In Progress source snapshot from SourceForge, but they will need to apply the following patch to the snapshot for maximum safety. I've tested this feature by creating a filesystem that had the backed up journal inode, and then trashing the relevant parts of the inode table. E2fsck used the backup journal information in the superblock to synthesize the journal inode, and then replayed the journal. The act of replying the journal replaced trashed portions of the critical metadata, and restored the filesystem back to its original state. I expect this should address the unrecoverable filesystem cases caused by a power failure causing the disk to scrible random crap on the inode table where the journal inode information is located. - Ted diff -Nru a/e2fsck/ChangeLog b/e2fsck/ChangeLog --- a/e2fsck/ChangeLog Thu Aug 21 12:00:49 2003 +++ b/e2fsck/ChangeLog Thu Aug 21 12:00:49 2003 @@ -1,3 +1,9 @@ +2003-08-21 Theodore Ts'o <tytso@xxxxxxx> + + * journal.c (e2fsck_move_ext3_journal): When moving the journal or + backing up the journal inode, make sure it gets written to + all the superblocks. + 2003-08-20 Theodore Ts'o <tytso@xxxxxxx> * journal.c (e2fsck_move_ext3_journal): If the superblock's backup diff -Nru a/e2fsck/journal.c b/e2fsck/journal.c --- a/e2fsck/journal.c Thu Aug 21 12:00:49 2003 +++ b/e2fsck/journal.c Thu Aug 21 12:00:49 2003 @@ -865,6 +865,7 @@ sb->s_jnl_blocks[16] = inode.i_size; sb->s_jnl_backup_type = EXT3_JNL_BACKUP_BLOCKS; ext2fs_mark_super_dirty(fs); + fs->flags &= ~EXT2_FLAG_MASTER_SB_ONLY; } } @@ -921,6 +922,7 @@ goto err_out; sb->s_journal_inum = EXT2_JOURNAL_INO; ext2fs_mark_super_dirty(fs); + fs->flags &= ~EXT2_FLAG_MASTER_SB_ONLY; inode.i_links_count = 0; inode.i_dtime = time(0); if ((retval = ext2fs_write_inode(fs, ino, &inode)) != 0) _______________________________________________ Ext3-users@xxxxxxxxxx https://www.redhat.com/mailman/listinfo/ext3-users
