2010-06-05 14:46:00

by DanVolkman

[permalink] [raw]
Subject: Kernel 2.6.32-22-generic sata problems

SYNOPSES
==========
Kernel fails to boot when multiple SATA drives are connected. After a
day of working on the problem (changing cables, sata positions,
motherboards, CPUs, power supplies, etc.) I discovered that with only
the boot drive connected everything worked fine.

BACKGROUND
=============
I had a career in UNIX so it is common for me to use cpio(1), tar(1),
dump(1), and dd(1) to replicate disks but have not been able to do so
with two SATA disks connected using older kernels but with this new
kernel I could not even have the second disk connected.

PROBLEM APPEARED WHEN
======================
An Ubuntu kernel update stopped my computer from booting. I have no
backup computer so my description is sketchy because I was not willing
to use a pencil and write the lengthy terse information. Boot error
message was something like:

Gave up waiting for root device. Common problems:0f106d71e58
...
ALERT! /dev/disk/by-uuid/..... does not exist. Dropping to a shell!


Dan
[email protected]





Attachments:
fstab (829.00 B)
ver_linux (1.31 kB)
sysinfo (3.41 kB)
grub.cfg (4.17 kB)
Download all attachments

2010-06-06 07:49:31

by Robert Hancock

[permalink] [raw]
Subject: Re: Kernel 2.6.32-22-generic sata problems

On 06/05/2010 08:45 AM, DanVolkman wrote:
> SYNOPSES
> ==========
> Kernel fails to boot when multiple SATA drives are connected. After a
> day of working on the problem (changing cables, sata positions,
> motherboards, CPUs, power supplies, etc.) I discovered that with only
> the boot drive connected everything worked fine.
>
> BACKGROUND
> =============
> I had a career in UNIX so it is common for me to use cpio(1), tar(1),
> dump(1), and dd(1) to replicate disks but have not been able to do so
> with two SATA disks connected using older kernels but with this new
> kernel I could not even have the second disk connected.
>
> PROBLEM APPEARED WHEN
> ======================
> An Ubuntu kernel update stopped my computer from booting. I have no
> backup computer so my description is sketchy because I was not willing
> to use a pencil and write the lengthy terse information. Boot error
> message was something like:
>
> Gave up waiting for root device. Common problems:0f106d71e58
> ...
> ALERT! /dev/disk/by-uuid/..... does not exist. Dropping to a shell!

Can you get any of the libata messages showing up on the console? You
might need to change grub options to boot with more debug displayed
(don't know how to do that offhand on Ubuntu).

Subject: Re: Kernel 2.6.32-22-generic sata problems

El Sun, 06 Jun 2010 01:49:27 -0600
Robert Hancock <[email protected]> escribió:


>
> Can you get any of the libata messages showing up on the console? You
> might need to change grub options to boot with more debug displayed
> (don't know how to do that offhand on Ubuntu).

Edit /etc/default/grub change
GRUB_CMDLINE_LINUX_DEFAULT="quiet splash"
to
GRUB_CMDLINE_LINUX_DEFAULT=""
Then run « sudo update-grub »

To see normal boot messages.

2010-06-07 18:34:44

by DanVolkman

[permalink] [raw]
Subject: Re: Kernel 2.6.32-22-generic sata problems


Thank you guys for responding to me email.

I did what you asked and it was nice to be able to see what messages
were being generated. In my debugging efforts I made one mistake, I
replaced the motherboard with the exact same brand/model motherboard
(different firmware though), I wish I would have had a different one
around the house. Anyway, it appears that the one month old drive I had
mounted failed and happened to coincide with the OS upgrade. I was not
using that drive and something seems suspicious because this is the
second new drive that has failed and these are drives from a reputable
company. With the changes in place you recommended I saw:

ata3.01: status: { DRDY ERR }
ata3.01: error: { UNC }
ata3.01: configured for UDMA/133
ata3: EH complete

This code would loop with occasional other messages. I unhooked my good
drive and used the Ubuntu distribution CD and booted up into "try
Ubuntu" mode and used the disk utility and it said the drive was bad. I
am not sure if this is a drive problem or a motherboard problem, maybe I
will replace both. Yes, I did try different SATA cables and sockets.

Anyway, sorry to bother you guys but you did help me identify the problem.

Dan
[email protected]


On 10-06-06 12:49 AM, Robert Hancock wrote:
> On 06/05/2010 08:45 AM, DanVolkman wrote:
>> SYNOPSES
>> ==========
>> Kernel fails to boot when multiple SATA drives are connected. After a
>> day of working on the problem (changing cables, sata positions,
>> motherboards, CPUs, power supplies, etc.) I discovered that with only
>> the boot drive connected everything worked fine.
>>
>> BACKGROUND
>> =============
>> I had a career in UNIX so it is common for me to use cpio(1), tar(1),
>> dump(1), and dd(1) to replicate disks but have not been able to do so
>> with two SATA disks connected using older kernels but with this new
>> kernel I could not even have the second disk connected.
>>
>> PROBLEM APPEARED WHEN
>> ======================
>> An Ubuntu kernel update stopped my computer from booting. I have no
>> backup computer so my description is sketchy because I was not willing
>> to use a pencil and write the lengthy terse information. Boot error
>> message was something like:
>>
>> Gave up waiting for root device. Common problems:0f106d71e58
>> ...
>> ALERT! /dev/disk/by-uuid/..... does not exist. Dropping to a shell!
>
> Can you get any of the libata messages showing up on the console? You
> might need to change grub options to boot with more debug displayed
> (don't know how to do that offhand on Ubuntu).
>

2010-06-07 23:35:33

by Robert Hancock

[permalink] [raw]
Subject: Re: Kernel 2.6.32-22-generic sata problems

On 06/07/2010 12:34 PM, DanVolkman wrote:
>
> Thank you guys for responding to me email.
>
> I did what you asked and it was nice to be able to see what messages
> were being generated. In my debugging efforts I made one mistake, I
> replaced the motherboard with the exact same brand/model motherboard
> (different firmware though), I wish I would have had a different one
> around the house. Anyway, it appears that the one month old drive I had
> mounted failed and happened to coincide with the OS upgrade. I was not
> using that drive and something seems suspicious because this is the
> second new drive that has failed and these are drives from a reputable
> company. With the changes in place you recommended I saw:
>
> ata3.01: status: { DRDY ERR }
> ata3.01: error: { UNC }
> ata3.01: configured for UDMA/133
> ata3: EH complete
>
> This code would loop with occasional other messages. I unhooked my good
> drive and used the Ubuntu distribution CD and booted up into "try
> Ubuntu" mode and used the disk utility and it said the drive was bad. I
> am not sure if this is a drive problem or a motherboard problem, maybe I
> will replace both. Yes, I did try different SATA cables and sockets.

Probably not a motherboard problem, but could be a power supply issue -
or maybe the drive is overheating causing premature failure.

>
> Anyway, sorry to bother you guys but you did help me identify the problem.
>
> Dan
> [email protected]
>
>
> On 10-06-06 12:49 AM, Robert Hancock wrote:
>> On 06/05/2010 08:45 AM, DanVolkman wrote:
>>> SYNOPSES
>>> ==========
>>> Kernel fails to boot when multiple SATA drives are connected. After a
>>> day of working on the problem (changing cables, sata positions,
>>> motherboards, CPUs, power supplies, etc.) I discovered that with only
>>> the boot drive connected everything worked fine.
>>>
>>> BACKGROUND
>>> =============
>>> I had a career in UNIX so it is common for me to use cpio(1), tar(1),
>>> dump(1), and dd(1) to replicate disks but have not been able to do so
>>> with two SATA disks connected using older kernels but with this new
>>> kernel I could not even have the second disk connected.
>>>
>>> PROBLEM APPEARED WHEN
>>> ======================
>>> An Ubuntu kernel update stopped my computer from booting. I have no
>>> backup computer so my description is sketchy because I was not willing
>>> to use a pencil and write the lengthy terse information. Boot error
>>> message was something like:
>>>
>>> Gave up waiting for root device. Common problems:0f106d71e58
>>> ...
>>> ALERT! /dev/disk/by-uuid/..... does not exist. Dropping to a shell!
>>
>> Can you get any of the libata messages showing up on the console? You
>> might need to change grub options to boot with more debug displayed
>> (don't know how to do that offhand on Ubuntu).
>>
>