[mythtv-users] Bus Error? Hard drive failing?

[Nick]

On 5/9/05, James Pifer <jep at obrien-pifer.com> wrote:
> I had a strange thing happen on my mythbackend. It's running FC3 (kernel
> 2.6.10-1.737_FC3) with 2 Maxtor 160gig drives in a logical volume.
> Sometime early Saturday morning, between 3 and 4 AM, the machine seemed
> to die. I already had a terminal session open on my laptop and if I
> tried to do anything I kept getting errors, like bus error. Nothing more
> specific than that.
> 
> I was stuck so I powered off and back on and all I got was "GRUB" in the
> top left corner and it would go no farther.
> 
> So I started playing a little. I booted from linux rescue CD and ran
> run_qtparted. It could see both hard drives, but it said /dev/hda was
> empty. It could see ext2 on /dev/hdb.
> 
> Then I booted from FC3 disc one because I wanted to see if the install
> would see the Linux installation. All it seemed to see was /dev/hdb.
> 
> Next I booted from Maxtor's drive utility floppy to check out the
> drives. I ran the quick test and long test on both drives and they both
> passed all the tests.
> 
> Then I rebooted the machine again, and now FC3 boots up! I had to go
> through fsck a couple times, but now it's working. It's made three
> recordings since then. The only problem I've had is when I tried
> watching a recording from a remote frontend I got this error on the
> backend a few minutes into it:
> WriteBlock(): Aborting WriteBlock, write to socket failed!
> 
> This machine has been very stable for a pretty long time. I did just add
> a Via VT6122 gigabit NIC last week. Up to this point it caused no issues
> in the machine, but the same NIC did seem to cause some instability in
> another FC3 machine I have. I don't know if it has anything to do with
> it or not.
> 
> Could this be the hard drive failing even though drive diagnostics says
> it's fine?
> Anything else possibly happening?
> 
> I googled for bus error but it seems to be a non-specific error, or at
> least I got results for all kinds of things causing bus error.
> 
> Any wisdom is appreciate!

No wisdom but a couple of suggestions:

i) have you got smartd running? If so, take a look at the SMART status
of the 2 drives and see if any errors have been recorded,
ii) check to see if the NIC is sharing IRQs with the drive controller
etc. Maybe change slot to try and get the NIC its own IRQ if it's
currently sharing, and
iii) remove/swap out the GBit NIC and see if the system's stability returns.

Nick