on Thu, Apr 18, 2002, hanasaki (hanasaki@hanaden.com) wrote:
Henrique de Moraes Holschuh wrote:
On Thu, 18 Apr 2002, hanasaki wrote:
had. Will I have problems becuase the newer card, or other vendor's
card, lays out the disks and / or data and does checksums
differently?
Yes.
If so, how do I mitigate this risk?
Either use software RAID, or buy two controllers. When one fails, use
the other to read the disks and move the data to another storage
system :)
Hmm and when the 2nd controller dies? Or if its doa when I need it?
Loooks like software Raid and a dual cpu ::(
I think you misunderstand RAID.
What is your risk scenario? What risks are you attempting to mitigate
via RAID?
What most RAID solutions provides you is the ability to ride through
single (or, with hot spares, multiple) disk failures, _allowing you to
schedule downtime for replacement at a time that's convenient_.
(Hot-swap systems allow you to bypass downtime).
RAID _doesn't_ guarantee data recovery from a fully hosed array. That's
what you need backups for.
RAID _doesn't_ guarantee system availability in the event of
catastrophic failure (think fire, flood, earthquake, plane crashing into
your building, sabotage). That's what redundancy and geographic
separation are for.
There are various risks that affect operations. There are various
solutions that mitigate risks. These come with both up front and
time-of-incident costs (including ease of recovery, downtime, data or
system loss, etc.).
Seems to me you've got a question/answer mismatch.
Peace.