Want ability to restore from failed upgrade.

Chris W. Parker cparker at swatgear.com
Wed Feb 23 20:30:17 UTC 2005


Jason Dixon <mailto:jason at dixongroup.net>
    on Wednesday, February 23, 2005 12:05 PM said:

> Stop right there.  Yes, you should have backups.  But don't rely on
> those solely as your backup in case of upgrade failure.  Your focus
> should be on building a test system and performing the upgrade there.
> Document and test everything.  Then proceed with your production
> system upgrade, referring to your documentation and, in the worst case
> scenario, restoring from the backups.

Yes great idea!

But this would require that I have another computer available with an
exact copy of my current filesystem on the live server would it not? And
that's what I don't know how to do. If I could configure an identical
system to what I've got running live that would be awesome because then
as you pointed out I could test everything and take notes and then just
restore if something went wrong and start over.

But I don't know how to make a copy of an fs nor do I know if this is
even necessary for a test system. I'd think that would it be only
because if I built a new system from scratch, using the same versions of
software that is on my live server, it would *still* be configured
different than my live server (i.e. configuration file changes long
forgotten about that would not be present on the test system).

What angle should I come at this from?



Thanks!
Chris.




More information about the redhat-list mailing list