[BBLISA] disk corruption recovery ideas?

[Eric Smith]

On 10/4/05, John Stoffel <john at stoffel.org> wrote:
>
> Eric> A few months ago I switched my home system's two external disks
> Eric> from USB 2 to FireWire hoping for a speed improvement (which
> Eric> didn't come.)
>
> Make sure the firmware on your external disk enclosures is upto date.
> I had horrible problems with a Prolific Technologies chipset in my
> USB/Firewire external enclosure until I upgraded it.  It was unstable
> under both Linux and Windows.  I was really disappointed.

I wanted to second John's comment about how bad the external cases are
that use the Prolific chipset.  I have 2 that worked well on USB 2
(just not very fast) but the FireWire implementation was... well,
non-functional.  I tried every firmware for that chipset and while
they got better they never actually worked correctly.  So if any of
you buy an external enclosure, be very careful about which chipset the
external enclosure uses.

Personally, I think that firmware issues caused my current problems. 
As a bit of new info, I moved the disks which had the errors over to
USB2 (fearing the issue could be FireWire related) and I got the exact
same error when I went to copy some files off the disk - just not on
the $MFT.  So this leads me to believe that the problem is with the
firmware in the enclosure.  I haven't turned them on since.  :)

I have now borrowed a system with 600G of free disk and I'm going to
image the drives with dd so I don't loose anything else.

>
> Eric> Once I did the switched I've had a occasional problem.... When I
> Eric> print (via USB2 on the motherboard) I have occasionally gotten a
> Eric> delayed write failure from Windows. It happens when I am
> Eric> printing and then tried to save a file from photoshop (to an
> Eric> external disk.) I'd just save the file again and everything was
> Eric> happy.
>
> So how are you doing backups of your data?  What are you going to do
> when a disk dies completely?  Hopefully, you're using some sort of
> RAID here...

I have a second disk that is a mirror of the first.  The mirroring is
via scripting and an rsync-like program for windows.  There is no RAID
here, but what I'm thinking of doing now will be RAID 5.  More on that
later.

> Can you get another IDE controller to put into the system (such as
> HPT302 or 3ware) and use that to hold your disks?  Or if you want more
> speed, you might think aboutt getting another USB2 PCI controller and
> hanging all the disks off that.

I could put another controller into my system.  My case does have the
capacity to hold those disks and I have at least one PCI slot free
(probably more.)

>
> Eric> The hardware facts:
> Eric> Win XP Pro with Service pack 2.
> Eric> Athlon XP 2000+ box, 1G ram 1 120G internal disk.
> Eric> 3 external disks (1 USB, 2 FireWire) all 250 Western Digital EIDE 8MB
> Eric> Cache "special edition drives".
> Eric> all onboard USB 2
> Eric> Adaptec PCI FireWire controller.
> Eric> Epson 4000 via USB2
>
> I assume you have a CDROM/DVD/DVD-RW or something else in there?  I'd
> only puy on disk/cdrom per-controller.  I never have Masters/Slaves on
> the same IDE bus, it just kills performance.

Yes, I have a CDRW and a separate DVD reader.  I didn't include them
only because they were not directly relevant to the problem (as
opposed to being involved in the solution.)  I fully agree with you
about master/slave issues with such devices.  My current setup has the
120G drive on its own channel and the two optical drives on the other.

> It might make sense to get a new case that has more drive bays and
> bring your disks internally, along with a new controller card.  That's
> probably what I would do.
> ...
> I've been thinking about getting a Network NAS server for home,
> something quiet and reliable.  I'm not as interested in speed as I am

What I haven't decided yet is if its just as easy to build a linux box
with an SATA controller, 4 250G SATA drives, RAID 5 them and export
them via Samba to my windows box over Gig ethernet.  Using lvm and
reiserfs I gain a lot of flexibility with disk layout.  I know others
who have done this and its basically a no-maintenance solution for
them (leveraging SMART drive monitoring to warn of impending disk
failure.)

The flaw I see in doing this is that I believe I either need tape
backup or another equivalent setup (physical second RAID5) to rsync to
to prevent bit corruption (not hardware failure.)  In other words, the
reason my current solution failed was because the backup disk was
accessible at the same time as the original, and the thing that
whacked one whacked the other.  I was protected against hardware
failure and it didn't fail.

> in reliability.  And doing backups, but my DLT7000 has flaked out on
> me... anyone know a good place to get it fixed?  It locks up the SCSI
> bus when it hits EOT... not good.

All the times I've had to deal with that class of tape drive, I've had
contracts that handle the repair.  So I can't help you there in
return.

Eric