>I do a full dump every night of the week. ...
OK, then you should be in good shape.
>I don't have the index files. Could you elaborate on the syntax. I am
>using dump, not tar, and my archives are compressed. I am not sure what
>you mean by $TAPE and i am a bit shaky on 'the-restore-program'. Also,
>where do I run this? The machine I want to restore from?
If the client and tape server machines have the same restore program
(e.g. they are the same OS type), you can run it either place. You may
run things on the tape server and then push the recovered file back over
to the client, or do everything from the client (which you'll have to
do if the OS's are different).
By $TAPE I meant the tape device name, which is probably something like
/dev/nst0 or /dev/rmt/0mn, etc. It depends on your OS. Take a look at
your amanda.conf. Unless you have a tape changer, it's whatever tapedev
is set to.
Let's assume you're going to run this from the client rather than the
tape server. First, cd to where you want to do the recovery. If you
want to restore the file directly in place, cd to the disk name you told
Amanda to back up. If you want to restore it someplace else and then
move it into place (which is what I **always** do, just to make sure
something bad doesn't happen), create a temporary directory in the same
disk and cd to there.
Then it would look like this:
rsh <tape-server> -n /path/to/amrestore -p <tape-device> <client> <disk> \
| gunzip \
| /path/to/restore ivbf 2 -
where:
<tape-server> host name of your Amanda tape server
/path/to/amrestore path to the amrestore program on the server
<tape-device> name of the tape device on the server
<client> host name of the client as listed in disklist
<disk> disk name of the client disk as listed in disklist
(this is a regular expression, see the amrestore
man page)
/path/to/restore path to the restore program that matches "dump"
The "restore" program that goes along with "dump" has an interactive mode
(that's what the 'i' flag gets you above). Once you get a prompt from it,
you can use "cd" to move around and then "add <filename>" to pick the file
you want and "extract" to start the extraction process. After the file
is recovered, you can "quit" to get out of "restore" and then you should
have a directory tree in your current working directory that matches the
path you recovered, with the file down at the bottom. If you were in the
top level of the backed up disk, you're done. Otherwise you'll need to
"mv" the file into the right place after making sure it's the right one
(e.g. the right version or whatever).
When this is all over, you might want to get a book about backups, or at
least read through the Amanda chapter:
http://www.backupcentral.com/amanda.html
>brian
John R. Jackson, Technical Software Specialist, [EMAIL PROTECTED]
