[mythtv-users] What is wrong with my drive?

[Kevin Kuphal]

On Thu, Jun 19, 2008 at 2:05 PM, Johnny Russ <jruss at mit.edu> wrote:

> Let me give a little background on how I got into my current conundrum. I
> got a couple of 250 GB SATA drives from a friend and I wanted to put them in
> my backend for extra storage. My backend is an old PIII 1 GHz Dell computer
> with two 160 GB PATA drives. I had to use a SATA to PCI card. I did a clean
> install and was using all of the storage in a RAID/LVM volume. I finally got
> everything set up and going but my system kept crashing. I tracked it down
> to some kernel bug with SATA to PCI cards that has been around for a while
> and is unlikely to be fixed soon. So I took out the 250 GB drives and figure
> I will put them to good use elsewhere. I have my system all backed up so I
> will just get it back to how it was. Simple right? :).
>
> I got everything set back up on the two 160 GB drives. My system started
> crashing again, but in a different way. It won't reboot (invalid disk
> error). I get things up with system rescue CD. One of the drives indicates
> there is no parition structure and gparted indicates that it is a 2048 TiB
> drive (I wish ;) ). Fdisk says it is -2042.... MB (yes that is a minus).
> This must be some default largest possible value I am guessing. At first I
> assume my drive is hosed. I run testdisk to see if it can find the
> partitions. It is giving me read errors on every cylinder as it is
> searching, so I back out of that with out finishing the search or making any
> changes. Voila! now gparted can see the drive fine with the correct size and
> partitions. I run fsck.xfs/e2fsck and everything is fine.
>
> The next day things crash exactly the same way. This time I realize the
> drives are connected to my motherboard in a different order than they were
> before. I have an old motherboard that has the 137 GB LBA limit issues. So I
> put the drives in the other order and copy my /boot and / partitions over to
> the good drive how things were before this whole mess started. I thought
> there might be an issue with the 137 GB limit. I remember previously that I
> was OK as long as I don't have any partitions larger than 137 GB. I set up
> things again but I split the storage up in to smaller partitions and then
> put them all into one logical volume with LVM. This time system doesn't
> crash anymore because the system files are on the good drive. But the LV
> goes down eventually and the bad drive is showing up with no partitions and
> 2042 TiB. Searching for a minute or two with testdisk and then quiting will
> frequently bring the drive back from oblivion. I don't know why.
>
> I ran smartctl --test=long and badblocks -svw on the drive and they came up
> with nothing. So I don't think the drive is actually faling. At least not in
> any normal way that I am familiar with.
>
> Smart indicates that it passes it self test when I tell it to run while the
> drive is in the messed up state. But  it does complain about some Smart data
> structures and the drive information is garbled. Here is what smartctl -a
> gives me when things are messed up:
>
> Warning! Drive Identity Structure error: invalid SMART checksum.
> === START OF INFORMATION SECTION ===
> Device Model:     Ma|tor$6]160T0$ $ $ $ $ $ $ $ $ $ $ $ $
> Serial Number:    ]45L\HWE$ $ $ $ $ $
> Firmware Version: ]AV45BW0
>
> When I bring the drive back to life by searching with testdisk, the smart
> data is back:
>
> === START OF INFORMATION SECTION ===
> Model Family:     Maxtor DiamondMax Plus 9 family
> Device Model:     Maxtor 6Y160P0
> Serial Number:    Y41LXHWE
> Firmware Version: YAR41BW0
>
> I realize this is a little OT but I haven't seen anything like this, and I
> know some people around here have probably put large drives into old
> systems. Any ideas what is going on or what I can do about it?
>
> I would rather not have to explain to my wife why MythTV is broken again
> because I was 'fixing' things.


Try a different IDE cable perhaps.  Maybe you bent/frayed or otherwise
mangled the ribbon in the messing around.  Check the pins all around to be
sure none are bent also.

Kevin
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://mythtv.org/pipermail/mythtv-users/attachments/20080619/3b6e6a00/attachment.htm