Re: [cobalt-developers] Restoring a RAQ4r

[Tim <timothy_bissell@xxxxxxxx>]

I understand both methodologies on the restore, the sites is so that when the
backup goes on, it retains the correct structures for the sites, instead of
crisscrossing domains. The reason for entering the users in such a manner, is
that, ideally the backup should keep the original permissions on the files
archived. And when it restores, it will reference ownership to uid and gid.

As for which directories to sync, off the top of my head, I would grab /etc,
/home (which has a symlink from /home/log/httpd) and on a raq3, it also has the
postgres database (/home/pgsql). If you need the old logs for some reason, grab
/var/log. And for the users mail, grab /var/spool/mail. I believe that should
cover it. let me know if you know of something else that should go in.


>
> > The tech person was very adamant about setting up all domain names and
> > users EXACTLY in the SAME ORDER as I had done the first time.  This seems
>
> (1) If you do ls -l /home/sites and look at the links
>      You can see the site # for each domain and they do go in numerical order.
> (2) Users I don't know.
>
> I am attempting to resolve the issue by using rsync to save the files
> once a day to a remote server (another RaQ). I am not sure which files
> to sync, I am currently syncing / , as if nothing else I can resync
> the restored server from these files.
> The problem with synching / is I am getting a lot of files that change
> as the system operates. I add --excludes to attempt to cull out the
> files I know I won't need.
> Can shed some light on which directories to sync???
> Gerald

--
Rule #1 - The customer is always right.
Rule #2 - They must pay for their arrogance!

Regards,
Timothy Bissell
Sun Microsystems Sr. Help Desk Technician
Phone: 1-800-266-4378