HPUX fsdb[1m]
fsdb(1M) fsdb(1M)
NAME
fsdb - file system debugger
SYNOPSIS
/etc/fsdb special [-b block#] [-]
Remarks
Always execute fsck(1M) after having running fsdb.
DESCRIPTION
fsdb can be used to patch up a damaged file system after a crash. It
normally uses the first super block for the file system located at the
beginning of the disk section as the effective super block. If the -b
flag is used, the block specified immediately after the flag is used
as the super block for the file system. An alternate super block will
always be found at block ((SBSIZE + BBSIZE)/DEV_BSIZE), typically
block 16.
fsdb deals with the file system in terms of block fragments, which are
the unit of addressing in the file system and the minimum unit of
space allocation. To avoid possible confusion, fragment is used to
mean that, and block is reserved for the larger true block. fsdb has
conversions to translate fragment numbers and i-numbers into their
corresponding disk addresses. Also included are mnemonic offsets to
access different parts of an inode. These greatly simplify the
process of correcting control block entries or descending the file
system tree.
fsdb contains several error-checking routines to verify inode and
fragment addresses. These can be disabled if necessary by invoking
fsdb with the optional - argument, or by the use of the O symbol.
Numbers are considered decimal by default. Octal numbers must be
prefixed with a zero. Hexadecimal numbers must be prefixed with 0x.
During any assignment operation, numbers are checked for a possible
truncation error due to a size mismatch between source and
destination.
fsdb reads a fragment at a time. A buffer management routine is used
to retain commonly used fragments of data in order to reduce the
number of read system calls. All assignment operations result in an
immediate write-through of the corresponding fragment.
The following symbols are recognized by fsdb:
# absolute address
i convert from i-number to inode address (for continuation
inodes as well as primary inodes; see inode(4))
b convert from fragment number to disk address
(historically "block")
d directory slot offset
Hewlett-Packard Company - 1 - HP-UX Release 9.0: August 1992
fsdb(1M) fsdb(1M)
+,- address arithmetic
q quit
>,< save, restore an address
= numerical assignment
=+ incremental assignment
=- decremental assignment
=" character string assignment
X hexadecimal flip flop
O error checking flip flop
p general print facilities
f file print facility
B byte mode
W word mode
D double-word mode
! escape to shell
The print facilities generate a formatted output in various styles.
Octal numbers are prefixed with a zero. Hexadecimal numbers are
prefixed with 0x. The current address is normalized to an appropriate
boundary before printing begins. It advances with the printing and is
left at the address of the last item printed. The output can be
terminated at any time by typing the interrupt character. If a number
follows the p symbol, that many entries are printed. A check is made
to detect fragment boundary overflows since logically sequential
blocks are generally not physically sequential. If a count of zero is
used, all entries to the end of the current fragment are printed. The
print options available are:
i print as inodes (primary or continuation)
d print as directories
o print as octal words
x print as hexadecimal words
e print as decimal words
c print as characters
b print as octal bytes
The f symbol is used to print data fragments associated with the
current inode. If followed by a number, that fragment of the file is
printed. (Fragments are numbered from zero). The desired print
option letter follows the fragment number, if present, or the f
symbol. This print facility works for small as well as large files
except for special files such as fifos, and device special files.
Dots, tabs, and spaces can be used as function delimiters but are not
necessary. A line with just a new-line character increments the
current address by the size of the data type last printed. That is,
the address is set to the next byte, word, double word, directory
entry, or inode, allowing the user to step through a region of a file
system. Information is printed in a format appropriate to the data
type. Bytes, words, and double words are displayed with the octal
(hexadecimal if X toggle is used) address followed by the value in
octal (hexadecimal if X toggle is used) and decimal. A .B or .D is
Hewlett-Packard Company - 2 - HP-UX Release 9.0: August 1992
fsdb(1M) fsdb(1M)
appended to the address for byte and double-word values, respectively.
Directories are printed as a directory slot offset followed by the
decimal i-number and the character representation of the entry name.
Inodes are printed with labeled fields describing each element.
The following mnemonics are used for inode examination and refer to
the current working inode:
a# data block numbers (0 - 14)
at time last accessed
ci continuation inode number
cno cnode ID for a device file
ct last time inode changed
gid group ID number
ln link count
md mode
mt time last modified
maj major device number
min minor device number
sz file size in byte unit
uid user ID number
The following mnemonics are used for directory examination:
di i-number of the associated directory entry
nm name of the associated directory entry
EXAMPLES
386i prints i-number 386 in an inode format. This now
becomes the current working inode.
ln=4 changes the link count for the working inode to 4.
ln=+1 increments the link count by 1.
fc prints in ASCII fragment zero of the file associated
with the working inode.
2i.fd prints the first fragment-size piece of directory
entries for the root inode of this file system.
d5i.fc changes the current inode to that associated with the
fifth directory entry (numbered from zero) found from
the above command. The first fragment's worth of
bytes of the file are then printed in ASCII.
1b.px prints the first fragment of the superblock of this
file system in hexadecimal.
2i.a0b.d7=3 changes the i-number for the seventh directory slot in
the root directory to 3. This example also shows how
several operations can be combined on one command
line.
Hewlett-Packard Company - 3 - HP-UX Release 9.0: August 1992
fsdb(1M) fsdb(1M)
d7.nm="name" changes the name field in the directory slot to the
given string. Quotes are optional if the first
character of the name field is alphabetic.
a2b.p0d prints the third fragment of the current inode as
directory entries.
WARNINGS
Use of fsdb should be limited to experienced fsdb users. Failure to
understand fully the usage of fsdb and the file system's internal
organization can lead to complete destruction of the file system and
total loss of data.
AUTHOR
fsdb was developed by HP and AT&T.
SEE ALSO
dumpfs(1M), fsck(1M), stat(2), dir(4), fs(4).
Hewlett-Packard Company - 4 - HP-UX Release 9.0: August 1992