On Fri, 22 Jun 2007 16:14:56 -0700, James Newtons Massmind wrote:
>> You might think backing up 5 workstations every hour would be time
>> consuming and network intensive but because XXCopy has some extensive
>> differencing options the whole process only takes 6-8 minutes per hour.
>>
>> It's awfully nice having automatic backups of everything back to the
>> last hour, day, week or month when you have one of those "oops" moments.
>> Obviously having a rotation of 4 weekly off-site backups is a good
>> thing too.
>>
>
> Does it have an option to store (not transmit, but store) only the
> difference between an original version of the file and the current
> version? E.g. if there is hugefile.dat and I copy it over once a month,
> then every day copy over hugefile.dat.200706##.dif so that I have the
> ability to restore to any point from the first to the last of the month
> and the space requirement is not sizeof(hugefile.dat)*31 ?

I don't believe so James. XXCopy is basically a file copy program on
steroids, HGH and crack combined. ;-)

However the program does have some hundreds of command line options so you
might want to download the user manual (which is refreshingly complete and
thorough, unusual for software these days!) and see if that feature might
be hidden away somewhere.

I don't have many (any?) huge data files that change frequently and my
backup storage requirements are fairly stable with the workstations and
servers, only adding a little data each day and an occasional new
application. I looked at the total sizes of the data stored on my systems
and doubled it to size my backup drives. They are currently at about 60%
usage. The one computer I do have that has massive data on it is an audio
workstation in my basement lab. I do audio editing on it and also
store/serve all my MP3 files there. I back it up separately to an external
USB drive.

I supplement the regular file based backups with periodic images of my
system drives using TrueImage. This allows a quick recovery from system
drive failure (it's happened a time or two over the years). Most of my
computers have at least two hard drives and I generally try to keep only
system related things on my "C" drive partition, despite what Microsoft and
other software vendors would have you do.

The reason I was drawn to XXCopy was that I wanted something that would
make an exact copy of the files and remove all orphan files of every drive
and directory on every system on the network. It does that really well.

Mostly I'm trying to protect against data loss due to hardware failure or
"oops" moments. So far my way has worked well even though it may not be
optimum. It is automatic and just quietly sits there and works and that was
also a major consideration for me.

My "servers" are really nothing more than stripped down installations of
Windows 2000 workstation. When configured like that I've found Win2K to be
*extremely* robust. My main server has over 6 months of uptime right now.
My workstations are configured similarly and about the only time I have to
reboot them is if I install software that insists on rebooting. My network
is a simple peer-peer network and it's all I need.

Every other backup program I've tried (Novaback and Retrospect mainly) had
problems for me.

Novaback, while easy to setup and use would only use it's own scheduler and
it would occasionally crash. It also puked it's database once it got above
about 50 GB of backup data. I'm not sure I ever got a reliable restore from
it once I started backing up my whole network.

Retrospect I found to be really cumbersome to setup and restore from. It
has extensive backup options (including several different ways of doing
differential backups). Again, I really wanted something that just kept a
separate copy of everything I do and can restore simply and quickly.

Frankly, I found commercial backup software to woefully inadequate and
unreliable. I bought and demo'd many packages and they all had problems.
Either they were aimed at single workstation backups or full domain
networks with 100's of workstations ($$$$) but finding anything that was
simple, robust and cost effective for a small workgroup was pretty much
non-existent. That's why I ended up rolling my own.

Matt Pobursky
Maximum Performance Systems

-- 
http://www.piclist.com PIC/SX FAQ & list archive
View/change your membership options at
http://mailman.mit.edu/mailman/listinfo/piclist