Raid 5 (was: Mounting of Linux volumes)

[Ross Werner]

On Fri, 2 Dec 2005, Gabriel Gunderson wrote:
> I've been reading this thread and wondering if you are talking about two
> different things.
>
> It seems that Nicholas is saying that an rsync is not enough to back up
> a server that has important and changing temporary data written to
> files.
>
> It seems that Ross wants to know if there is any harm in rsyncing the
> directories where that data lives.

What I'm wondering is this:

on a typical Linux server, if you just do an rsync of the entire drive at 
2 AM, and the live hard drive bites the dust next day, if you restore the 
dead drive from the rsync'd backup:

  a) What sort of things are not going to just magically work, and
  b) What sort of things are going to possibly have data loss because the 
drive was not read-only when rsync'd?

My own impression of (a) is that there are a lot of things that won't just 
magically work; you'll be lucky if you get a bootable system out of a raw 
file-copy. But reinstall Linux and possibly a few of the programs and 
you're good to go. I'm not so concerned about this one.

The answer to (b) is much more worrisome to me ... I understand that 
databases could have data loss because you're rsyncing in the middle of 
whatever the database happens to be doing at that particular moment. I 
solve this problem by dumping the database to a flat file and backing that 
up instead. But what other sorts of things might have data loss? For 
example, if a mail file was being written to at the exact moment when you 
rsync'd, and that caused the remote backup file to be deleted or zeroed 
out or corrupted or whatever.

I think we've established that mail files are ok because if you rsync 
while it's being written to, you'll simply get a backup of the old version 
of the file (not an empty or deleted file). But are there other programs 
where you *will* get data loss in this sort of scenario?

 	~ Ross