The smallest atom
dump can backup is a filesystem, so in your case
Code:
/dev/ad0s1a (/)
/dev/ad0s1g (/home)
/dev/ad0s1f (/tmp)
/dev/ad0s1d (/usr)
/dev/ad0s1e (/var)
By default
dump assumes to be working on a tape drive device. If you want to backup to a file you have to use both the
-f and
-a options.
You have to watch out that you don't backup the filesystem you backup device is mounted on.
E.g if you are dumping
/dev/ad0s1g, your
/home directory, don't mount your drive on a subdirectory of
/home.
If you backup drive is mounted on "/mnt", then it is ok to dump any other filesystem, except /dev/ad0s1a (/) of course
If you are the only user, you could mount your disk on a directory in /tmp, which you probably don't want to backup anyway.
Assuming your backup drive has been mounted on
/tmp/mount you could do something like this
Code:
# dump -0auf /tmp/mnt/myhome_dump0 /dev/ad0s1g
The options
Code:
0 : level of backup
a : auto-size to bypass tape length stuff
u : update dump statistics in /etc/dumpdates
f /tmp/mnt/myhome_dump0 : destination file
/dev/ad0s1g : the filesystem to be 'dump'ed
Or you could compress with
gzip
Code:
# dump -0auf - /dev/ad0s1g | gzip >/tmp/mnt/myhome_dump0.gz
Here you specify the destination file with a hyphen:
-.
This means standard output, which is piped into
gzip for compression.
All this works similar in FreeBSD and OpenBSD. Most system administrators will either go into single user mode, or stop daemons/programs which could generate file output during the dumping process.
With FreeBSD, you don't have to do this, if you the
-L option which takes a snapshot of the filesystem and dumps the snapshot.
But I am conservative, I just go into single user mode