Archived from groups: comp.sys.ibm.pc.hardware.storage (
More info?)
On Fri, 14 May 2004 01:23:25 GMT, "Ron Reaugh" <ron-reaugh@att.net>
wrote:
>> I was thinking of cheap small drives in cheap removeable trays. That
>> way if an entire drive goes south, you have other drives available.
>I'm not following. Are you talking about RAID 1 drives or backup drives
>with compressed image files?
In the statement above I was talking only about the RAID 1 part of the
overall configuration. I was contrasting your huge drives with my
desire to have a cheap system. Whereas you get multiple volumes on
those huge drives, you pay a price for it - first there is the cost of
the huge drives and there is the risk that the drive will die on you.
With multiple small cheap drives I would swap them in one after the
other creating multi-period disaster-recovery volumes. For example, if
I used a configuration that had 6 drives, I could put 2 in the RAID 1
system, and have 4 to rotate thru the mirror each week for 1 month.
One thing I like about that is you can pick and choose which week in
the past to restore from. For example, I got burnt badly a short while
ago by one of Microsoft's so-called hot patches. The damn thing really
messed up IE, which crashed an application that uses things from IE
extensively. I tried to back out the patch but no luck. I tried to
repair the Win2K operating system, including reinstalling SP4 but no
luck. So I restored the system with the "archive" I had made with
Drive Image Pro.
That "archive" was 6 weeks old which in this case was a good thing
because I had installed hot patches as far back as 3 weeks. By going
back longer than that I ensured I was getting a hot-patch-free
restore. My point is that if you keep 4 weeks of "archives", you can
decide to go back to any one of them based on the history of updates
along the way.
This whole scheme hinges on the ability to find a RAID 1 controller -
either PCI-based or mainboard-based, which will allow the use of an
"archive" for disaster recover. I am sure that all the controllers
allow for taking a drive out and rebuilding the mirror on a different
drive, thereby creating the "archive" drive on the shelf for later use
if necessary. The question seems to be whether you could then pull the
2 current RAID 1 drives and use the "archive" drive to start over.
Separately from all this RAID 1 configuration is the business of a
daily backup to a third drive bad with a very small drive using a
differential backup strategy. That is not the same as an incremental
backup because with the differential backup the archive bit is not
set. Therefore the same files get backed up each time. However you can
require that each backup uses the same backup file which means that
the size of that backup file will not get too large. Hence the ability
to use a very small drive. I am using a 4GB drive which has more than
enough capacity for my 300KB backup file - which uses less than
200KB because I compress it too.
I suppose the single most valuable thing I have discovered on this
forum is that there is no excuse for someone not to have a way to
restore his system no matter what happens. If I had been keeping daily
backups on that very small drive, then when I did the 6-week old
disaster recovery restore I would have been able to bring it up to
date because before I would have replaced the current drive contents
with the old "archive" I would have done a manual backup to flush
everything to a file.
What could be easier than all that?
--
Map Of The Vast Right Wing Conspiracy:
http://www.freewebs.com/vrwc/
"You can all go to hell, and I will go to Texas."
--David Crockett