On Mon, Dec 17, 2007 at 09:58:06PM -0800, Neil Schneider wrote:
David Brown wrote:
On Mon, Dec 17, 2007 at 09:56:37AM -0800, Tracy R Reed wrote:
David Brown wrote:
Ugh, I guess this means I'm back to using tar, and writing my own
backup software. I have filed a bug report with SGI, so we'll see
where that goes.
Have you looked at bacula? I've tried most other backup systems and bacula
is the only one I have stuck with for any length of time.
Bacula falls under the list of "doesn't understand what an incremental
backup is". It's on their "not implemented page", but as far as I'm
concerned makes it utterly worthless. What exactly is the point of backup
software if it can't restore data that was on my machine at the time of
backup.
Well that's interesting, because apparently either I don't understand what you
are saying or the web page is inaccurate. I have restored from bacula backups
and have gotten all the data, up to and including the most recent incremental.
I had a catastrophic system failure and had to restore from backup.
How do you know you restored correctly? Do you have a tool that verifies
that what you restore is the same as what was on the original machine? If
you've ever renamed a directory, it is likely you've got at least one file
that ended up in the wrong place.
I hope I don't just come across as arrogant about this, it's more just my
frustration. Very few programs that present themselves as backup programs
actually will restore the data backed up accurately. It's frustrating to
discover programs that have put lots of effort into scheduling and
distributing the backup, and seem to have neglected the part about
correctly backing up the data.
So who exactly says it won't restore correctly, besides you?
It's in the main webpage: Current State of Bacula, listed as a Current
Implementation Restriction. At least the Bacula developers understand the
problem, they just don't seem to think it's important.
I posted an example, try it, it fails most backup programs. There are more
advanced cases that fail even more programs. Something like xfsdump almost
always works, except in hard to reproduce situations, such as deleted
inodes getting reused as a different type of object.
Bacula's solution at least fails a little better than most. You get all of
your data, just not where you left it.
It's even more frustrating when commercial backup software can't get it
right. There's a whole slew of fairly cheap programs that are basically
some person implementing tar with some special added feature, except they
didn't get the tar part right.
Dave
--
[email protected]
http://www.kernel-panic.org/cgi-bin/mailman/listinfo/kplug-list
