Hello!
Thanks at:
Olle Eriksson
Tomasz Tyrakowski
Angela amw_at_aber.ac.uk
Larry Griffith
Andre Florath
Huehls, Mark R.
Richard Eisenman
Pedro J. Lobo
Javier Aida
alan_at_nabeth.cxo.dec.com
BOLSON
Eugene Chu
Marcelo Alonso M.
My priginal question was:
>We are using the utility vdump for to do backups of advfs partition and
>we have a 4mm DAT . We use the device /dev/rmt0h (compressed)
>for can to save more of 2Gigabytes but less that 4 Gigabytes.
>My advfs partition has 3,200 GigasBytes but it doesn't fit into the
>DAT, the vdump ask other DAT for to finish.
>My question is: Is posible that the size of the saveset is > 800Mbytes?
>or Why don't fit into the DAT?
Here is the answer:
alan_at_nabeth.cxo.dec.com
How well data compresses depends on the data. If the data
has lots of repetition, it will may compress far better than
2:1. If it contains little or no repetition it may compress
barely above 1:1 or even get larger. The problem of the data
getting larger is a common one when you try to compress data
which is already compressed. You can see the affect using
compress(1):
% compress -v < some-medium-sized-file > compressed.Z
% compress -v < compressed.Z > compressed.ZZ
% ls -l compressed.Z compressed.ZZ
If your file system contains enough data that is already
compressed, then it might be wise not to try and compress
it again. Also, vdump supports an option to do software
compression of the data. Using this in combination with
a tape drive that compression is likely to be a waste in
time and tape capacity.
To get a rough estimate of how well your data compresses you
can use:
vdump 0f - /file-system | compress -v | wc -c
Simply, do an ordinary backup, but write to stdout instead of
the tape and pipe the result to compress. Then pipe it to wc
to count the size of the data. If you get negitive compression
then your data probably won't fit on a single tape that holds
2 GB uncompressed.
Received on Mon Dec 15 1997 - 13:42:20 NZDT