[adr-devel] Essential Backup Tool

[Bernd Harmsen]

Hello Stefan,

On Wed, Apr 23, 2003 at 08:08:31PM +0200, Stefan G. Weichinger wrote:
> I restarted working on the Essential Backup Tool today.
Great.


> I want to reconstruct the former oops_backup.sh to deliver
> machine-readable information in a "VARIABLE=VALUE"-style. If you think
> another format would be better please tell me.

I am not shure in the moment. A file like this would be good, because
we can simply "source" it in the bash-script that does the disaster
recovery. All the interpretation of program outputs must than be done
in the essential backup script.

Maybe there are problems with putting multiline text-values in
bash-variables because the linefeed ist lost or must be quoted. (But I
am not realy shure).


> I also took out the "autosearch"-routines and changed the AM_CONF_DIR
> to be hardcoded as discussed.
Good.


> I also commented out the USB-stick code.
> I want to concentrate on content first, the delivery of the data comes
> second, I think.

Yes. I think this is another script. The essential backup script should
collect data, interpret it  and deliver it in an automated way.


> The system-dependent parts will be clearly determined and marked so
> maybe Bernd can put in the debian-specific parts. Or we merge Bernds
> script into.

Good. I will add the debian stuff.
We need a good test to know on which distribution or plattform we are
running. For the plattform we can maybe use "uname" and "uname -m" On
all Debian systems the is a file /etc/debian_version containing the
versions number of the distribution. 


> I will give you a list of variables the script generates, so you can
> think about it and add wishes ...

I think this is the most important step. If you send me a list I will
think about that and add. Please, consider that the script must run on
amanda backup-server and backup-clients.


> Tell me if that all is ok for you.
Sounds nice.


Regards,
Bernd