Unix |
Unix v6 |
|
![]() |
dump(5) |
![]() |
dump incremental dump tape format The and commands are
used to write and read incremental dump magnetic tapes. The
dump tape consists of blocks of 512-bytes each. The first
block has the following structure.
struct {
int |
isize; |
||
int |
fsize; |
||
int |
date[2]; |
||
int |
ddate[2]; |
||
int |
tsize; |
};
and are the corresponding values from the super block of the
dumped file system. (See file system (V).) is the date of
the dump. is the incremental dump date. The incremental dump
contains all files modified between and is the number of
blocks per reel. This block checksums to the octal value
031415. Next there are enough whole tape blocks to contain
one word per file of the dumped file system. This is divided
by 16 rounded to the next higher integer. The first word
corresponds to i-node 1, the second to i-node 2, and so
forth. If a word is zero, then the corresponding file
exists, but was not dumped. (Was not modified after If the
word is 1, the file does not exist. Other values for the
word indicate that the file was dumped and the value is one
more than the number of blocks it contains. The rest of the
tape contains for each dumped file a header block and the
data blocks from the file. The header contains an exact copy
of the i-node (see file system (V)) and also checksums to
031415. The next-to-last word of the block contains the tape
block number, to aid in (unimplemented) recovery after tape
errors. The number of data blocks per file is directly
specified by the control word for the file and indirectly
specified by the size in the i-node. If these numbers
differ, the file was dumped with a ‘phase
error’. dump (VIII), restor (VIII), file system(V)
![]() |
dump(5) | ![]() |