[Nottingham] Live system backup pitfalls / problems

[Martin]

David Bean wrote:
> This really goes for any backup...
> 
[... Some good stuff ...]
> 
> Why you ask?  What if more than one server goes down for any reason?
> (fire, hacking etc) You can't be expected to fix it all at once can you?

Yes! (:-|)

Sometimes one of reality and expectations have to be modified... All 
part of improved education and/or improved planning...


> Your boss will be much happier if you can enlist staff to get everything
> running in three hours than doing it all yourself in ten.

Yes with the proviso of whether you can be sure they understand what 
they are doing and have an _interest_ in what they are doing. Its far 
too easy for someone to abondon a slow backup/restore to rush off for 
lunch or to ignore an unexpected error message...


Slightly OT from what I was asking but too good not to mention is:
http://www.taobackup.com/

Can't remember where I stumbled across it but it is good if only for the 
thoughtful presentation... Enjoy.


Still open for comments on whether a live backup with various active 
system files and open sockets has any irksome 'gotchas' with it...

(Sockets don't get backed up as the one's I know of all get generated 
automatically.)


Further thoughts welcome,
Martin

-- 
----------------
Martin Lomas
martin at ml1.co.uk
----------------