|
Re: Define what holding disk to use?
2008-03-06 22:13:42
Aaron J. Grier wrote:
On Tue, Mar 04, 2008 at 08:29:03AM -0500, Chris Hoogendyk wrote:
My own preference is to configure the system with separate drives for
holding disk. There is no reason for them to be raid.
all backup data passes through the holding disk: doesn't it follow that
the holding disk should be as reliable as possible? and doesn't that
imply at least a mirrored holding disk? MTBF may be lowered, but
failure of a single holding disk will not affect backups.
Why?
It's transient data. It's backup on its way somewhere else. Get a fast,
high quality, reliable drive. Or two of them configured as two separate
holding disks to speed up parallel backups.
In about a decade with about a dozen servers and maybe 60 drives, we've
never had a catastrophic drive failure. We've had 2 drives that started
showing enough errors that we replaced them. We are currently buying
server class Seagate Cheetah Ultra320 SCSI drives.
If one backup in 10 years fails, frankly it's just not a big deal. Run
it over.
If, on the other hand, the disk *is* your backup, then RAID makes sense.
Do raid 5 with a hot spare. Monitor your logs so that you see a failure
coming as soon as possible. But I really prefer the multiple redundancy
of a long tape cycle with several fulls and daily incrementals all on
individual tapes. If one tape "crashes", I've probably got several other
copies of the same data.
---------------
Chris Hoogendyk
-
O__ ---- Systems Administrator
c/ /'_ --- Biology & Geology Departments
(*) \(*) -- 140 Morrill Science Center
~~~~~~~~~~ - University of Massachusetts, Amherst
<hoogendyk AT bio.umass DOT edu>
---------------
Erdös 4
|
|
|
