problem with dump/restore across filesystems

[George Robbins]

In article <22152 at felix.UUCP> gordon at prls.UUCP (Gordon Vickers) writes:
> 
>     To aid in the speady customization of Ultrix2.0, prior to 
>  doing the installation, I decided to free up an RA80 for awhile
>  and just backup my system disk by copying one filesystem to another.
> 
>     For some reason, my backup filesystems are about 10Mbytes larger
>  than the originals.  Prior to doing the backup, I used 'newfs' and
>  then 'df' to verify that the new filesystem was empty (except for
>  an empty lost+found directory).

I'd check the parmameters used in /etc/disktab to newfs/mkfs the file
system.  Different partitions on the RA80/RA81 have different fragment
and block sizes defined, which would create different "waste" factors
when loading the filesystem.

There are also some classes of files, notably those managed by the 'dbm'
routines that are sparsely allocated and will expand to full size when
copied sequentially.  The most commom expample for a usenet system is
the news history file.

It's a bit confusing, you mention freeing up an RA80, but your information
shows you copying from a 'd' partition, while the /etc/disktab info for
the RA80 doesn't list a 'd' partition.

-----

For the Gentleman from Australia that asked about using a Fujitsu Eagle
with Ultrix - the Eagle should work with the standard ultrix drivers.
Depending on the emulation proms included with the controller and the
switch settings you can either emulate some arrangement of DEC drives
or use the eagle entry in /etc/fstab.  Check with Emulex to make sure...