Whether a device is hot swappable depends on two or three things:
1) If the electrical connection can be broken without transient (or the
bus can recover from such a transient)
2) If the hardware interface can handle attaching and detaching devices,
recognizing them in real time
3) (depends on the setup) if the software can handle the attachment and
detachment.

This means you need a hot-swap enclosure or connector for the drive, a
controller that understands what's going on, and possibly an OS that
knows what's going on.

There are hot-swap for both SCSI and IDE drives, and many hardware RAID
cards handle it transparantly to the system.  The Promise FastTrack, for
instance, with hard drive hot-swap bays in a mirror or mirror/stripe
configuration will allow you to take a drive out of the system while
running, replace it, and the FastTrack will rebuild the data on the new
drive from the current drives.  All this _can_ go on without the OS ever
knowing what's happening (but all RAID cards come with monitor programs
so the OS can know what's going on and take appropiate action (ie,
disable web input to limit disk writes until the drive is replaced-the
server still functions, but in a limited sort of way which prevents
further problems))

Many, if not most, RAID cards will handle this sort of situation.
Generally you'll spend about $40 to $100 for each drive bay (these are 5
1/4 units which hold 3.5 inch drives), and (depending on the bus and
model) you'll spend anywhere from $30 to $500 for a RAID card.

For your situation a simple RAID 0 (mirroring only) would suffice.  It
is especially good for web serving as RAID 0 reads twice as fast (since
it can read from both drives at once) than a single HD would (though
writing takes place at the same speed as the slowest drive in the RAID).

-Adam


James Newton wrote:
>
> Its back up. But I don't like this kind of problem. I'm going to concentrate
> on figuring out what I did wrong with the UPS to Server communication and
> then try to get together some hard drives to get RAID going.
>
> In a raid system, are standard drives "Hot Swappable"? I.e. can you just
> pull the cover off the box, unplug the drive that failed and replace it and
> then format the new drive and have NT replicate the data from the other two
> drives back onto it?
>
> ---
> James Newton (PICList Admin #3)
> mailto:jamesnewton@piclist.com 1-619-652-0593
> PIC/PICList FAQ: http://www.piclist.com or .org
>
> -----Original Message-----
> From: pic microcontroller discussion list
> [mailto:PICLIST@MITVMA.MIT.EDU]On Behalf Of j newton
> Sent: Tuesday, February 13, 2001 12:15
> To: PICLIST@MITVMA.MIT.EDU
> Subject: [OT]: PICList.com down for Hard Drive damage
> Importance: Low
>
> Well, thanks to the rolling blackouts in California, and something not
> working with the UPS, the server had an unscheduled, unwarned, power fail
> last night and apparently one Saturday as well that I didn't see.
>
> Again, my little script on the Linux box at the Escondido address did not
> notice that the main sever was not responding (I think it must be looking at
> a cashed page from my ISP and not actually making contact with the NT server
> it is supposed to be watching) and so it did not switch DNS to the backup
> server at the same location.
>
> So the NT server log shows a restart with no matching shut down. I'm trying
> to figure out why the cable from the UPS to the server that is supposed to
> tell NT to shut down when the power goes out and before the UPS gives up
> didn't work.
>
> I though that we were ok, but this morning I started getting "bad sector"
> type errors and realized that I would have to CHKDSK the entire thing again.
> After logging everyone out and running the tests, it would appear that some
> permanent damage was done to the hard drive. For now, the bad sectors have
> been marked and are not being used, but I'm concerned about the future of
> that drive.
>
> It's time to put two more drives in and setup RAID. I should also install a
> tape backup unit so that if the system volume dies it won't take 3 days to
> set up the OS again... The content all gets backed up nightly over the
> network to a set of CDRW drives but that won't work for system files. Unless
> you know something I don't.
>
> Anyway, I'm formatting the content volume and then restoring from backup so
> the server will be up again in a bit.
>
> Sorry for the inconvenience.
>
> ---
> James Newton (PICList Admin #3)
> mailto:jamesnewton@piclist.com 1-619-652-0593
> PIC/PICList FAQ: http://www.piclist.com or .org
>
> --
> http://www.piclist.com hint: The PICList is archived three different
> ways.  See http://www.piclist.com/#archives for details.
>
> --
> http://www.piclist.com hint: The PICList is archived three different
> ways.  See http://www.piclist.com/#archives for details.

--
http://www.piclist.com#nomail Going offline? Don't AutoReply us!
email listserv@mitvma.mit.edu with SET PICList DIGEST in the body