Lan Barnes wrote:
Over the weekend I installed a 6 G Quantum IDE HD as added space on my
SCSI server. I had no problem. I then moved about 2 G of digital photos
to it and deleted them from the original space. A mistake.
Was this a known good hard drive?
Later, I attempted to move about a gig of pix from the laptop to this
server through a wireless connection. Performance plunged, and I closed
the connection abruptly and rebooted the server.
I don't see how pulling the wireless connection's plug, so to speak,
would corrupt any but possibly the file currently being processed.
Several of these steps are now obvious mistakes.
As in the difference in safety between cp+rm vs. mv? ;-)
The server now refuses to come up with this HD on line. It gives a
message that there is a I/O buffer failure on /dev/hdd (the "new" drive)
while loading init, and refuses to boot past that, even when the drive
is removed from /etc/fstab and the BIOS. I have to take power off the
drive to boot up. Then everything is OK, except my photos are gone.
Are you sure the server's power supply is up to the added loaded of an
additional HD?
So what is recommended? I plan to try (in this order)
Remarks or advice?
Case Study: I recently had a problem with a couple of HD's fail on my
server. They had been running for about a year since purchase with no
problem. Both drives were identical 120 GB IDE types in an external SCSI
RAID 0 enclosure. That amounted to _all_ of my file server's 240 GB of
data capacity!
The short story is, the drives were fine, but the enclosure's PS was
failing. My solution was to power up the two drives from an ATX power
supply rather than the enclosure's [1] just long enough to copy all that
data to a new 250 GB drive.
I suggest that you download and run the drive mfgr's diagnostic software
before you go mucking with the drive yourself. Generally that
software boots from floppy or CD so you can disconnect power from any
other drives first, in case it really is a PS problem.
If the drive checks out okay then you have several obvious paths to
preserving its data. My guess is that even if the drive is physically
okay, you may well have some corrupted data as a result of whatever
non-HD-related source is the real problem.
If the drive is actually bad, the diag software may be able to get it
usable, but again, probably at the expense of /some/ lost data.
As I always recommend: do *not* jump to the conclusion that your HD is
bad just because data is being corrupted or even if it doesn't work.
It's not running in a vacuum. It depends on several other critical
components in your system to do its job reliably.
Another lesson that might or not be relevant here: don't put critical
data on a questionable and untested lump of hardware.
[1] I still needed the external box to run as it was providing the IDE
<-> SCSI bridging as well as the RAID support. But its own PS was
sufficient for those tasks.
--
Best Regards,
~DJA.
--
[email protected]
http://www.kernel-panic.org/cgi-bin/mailman/listinfo/kplug-list
