# Repairing filesystems with fsck/e2fsck

On April 12 a linux box experienced problems booting.  
The ultimate problem was that there were no filesystems listed in /etc/fstab.
On the way to discovering this we had to do some diagnostics to ensure that the filesystems were not corrupt.  
To do this, we use fsck for ext3 filesystems and e2fsck for ext2 filesystems (of which we do not have any).

Here are some steps that might be helpful when getting 'bad superblock' and filesystem errors on boot:

#1) Get to a command-line environment:
	o Drop into a 'repair shell' if it gives you the option.
	o Boot from the Knoppix CD if no other options exist.

#2) Check to see what filesystems are mounted:
	o Run # > df -h

#3) Check the filesystem mappings:
	o Run # > fdisk <device>
	o Here, <device> may be /dev/hda, /dev/hdb, /dev/hdd, or some other filesystem.
        o Here is the output from running # > fdisk /dev/hdb on Pecos...
========================================================================================
[root@machine_name username]# fdisk /dev/hdb
 
The number of cylinders for this disk is set to 30401.
There is nothing wrong with that, but this is larger than 1024,
and could in certain setups cause problems with:
1) software that runs at boot time (e.g., old versions of LILO)
2) booting and partitioning software from other OSs
   (e.g., DOS FDISK, OS/2 FDISK)
 
Command (m for help): p  #this prints the fdisk journal
 
Disk /dev/hdb: 250.0 GB, 250059350016 bytes
255 heads, 63 sectors/track, 30401 cylinders
Units = cylinders of 16065 * 512 = 8225280 bytes
 
   Device Boot    Start       End    Blocks   Id  System
/dev/hdb1   *         1        13    104391   83  Linux
/dev/hdb2            14      1288  10241437+  83  Linux
/dev/hdb3          1289      1670   3068415   82  Linux swap
/dev/hdb4          1671     30401 230781757+   f  Win95 Ext'd (LBA)
/dev/hdb5          1671      1925   2048256   83  Linux
/dev/hdb6          1926      2052   1020096   83  Linux
/dev/hdb7          2053     30401 227713311   83  Linux
 
Command (m for help): q  #this quits the fdisk environment
========================================================================================

#4) Check /etc/fstab to see what needs to be mounted:
	o Here is a copy of /etc/fstab file from Pecos...
========================================================================================
machine_name: ~>cat /etc/fstab
/dev/hdb5               /                       ext3    defaults        1 1
/dev/hdb1               /boot                   ext3    defaults        1 2
none                    /dev/pts                devpts  gid=5,mode=620  0 0
/dev/hdb7               /home                   ext3    defaults        1 2
none                    /proc                   proc    defaults        0 0
none                    /dev/shm                tmpfs   defaults        0 0
/dev/hdb6               /tmp                    ext3    defaults        1 2
/dev/hdb2               /usr                    ext3    defaults        1 2
/dev/hdb3               swap                    swap    defaults        0 0
/dev/hda2               /home2                  ext3    defaults        1 2
/dev/hdd1               /home3                  ext3    defaults        1 2
/dev/cdrom              /mnt/cdrom              udf,iso9660 noauto,owner,kudzu,ro 0 0
/dev/fd0                /mnt/floppy             auto    noauto,owner,kudzu 0 0
========================================================================================

#5) Check filesystem integrity on partitions to be mounted:
	o Run # > fsck /dev/hdb###
	o Here, ### should be replaced by the map number (1,2,3,5,6,7) !Not on /hdb4!
	o Running #fsck# should tell you the name of the partition and if it is clean.

#6) Repairing filesystems:
	o You can interactively repair filesystems using # > fsck -r <device>
	o You could use a different superblock using # > e2fsck -b 8193 <device>

#7) Mount the clean filesystems:
	o Run # > mount /dev/hdb1 /boot
	o Run # > mount /dev/hdb2 /usr
	o Run # > mount ...and so on...
	o Check to see if they have been mounted by running # > df

#8) If you are in a read-only environment, try this:
	o Run # > mount -o remount,rw <device> <mountpoint>

# Other possibilities:

	o Try running 'debugfs', it is an ext2 filesystem debugger:
========================================================================================
DEBUGFS(8)                                                          DEBUGFS(8)
 
NAME
       debugfs - ext2 file system debugger
 
SYNOPSIS
       debugfs [ -b blocksize ] [ -s superblock ] [ -f cmd_file ] [ -R request
       ] [ -V ] [ [ -w ] [ -c ] [ -i ] [ device ] ]
 
DESCRIPTION
       The debugfs program is an interactive file system debugger. It  can  be
       used to examine and change the state of an ext2 file system.
       device  is  the special file corresponding to the device containing the
       ext2 file system (e.g /dev/hdXX).
========================================================================================

	o Try running 'dumpe2fs' for block information 

========================================================================================
DUMPE2FS(8)                                                        DUMPE2FS(8)
 
NAME
       dumpe2fs - dump filesystem information
 
SYNOPSIS
       dumpe2fs [ -bfhixV ] [ -ob superblock ] [ -oB blocksize ] device
 
DESCRIPTION
       dumpe2fs  prints  the  super block and blocks group information for the
       filesystem present on device.
 
       dumpe2fs is similar to Berkeley’s dumpfs program for the BSD Fast  File
       System.
========================================================================================