WARNING: Kernel Errors Present

Alan Peery peery at io.com
Thu Jan 4 20:12:43 UTC 2007


Lyvim Xaphir wrote:
> On Thu, 2007-01-04 at 12:43 +0000, Anne Wilson wrote:
>   
>> Warnings like this crop up regularly in logwatch:
>>
>>     hdd: command error: error=0x54 { Ab ...:  4 Time(s)
>>     hdd: command error: status=0x51 { D ...:  4 Time(s)
>>     
> Are you on Core 6?
>
>   
I had similar errors on Core 6 working with SATA hard drives.  When the 
kernel noticed it would mark the related file system read only--which 
was a show-stopper as I was operating with two partitions striped across 
all four hard drives.  The fault was intermittent, and I could operate 
for hours before it occurred.  All new hardware with a proper sized 
power supply, but I tried cutting back on drives to cut power 
consumption.  It didn't help.  It wasn't temperature related, as it 
happened once very shortly after a full cool down.  It wasn't IO load 
related, it would fail regardless of how hard I was hitting the disks.

The mother board is a Asus PN5SLI.  For the flexibility I wanted 
(multi-boot with Windows, LVM striped filesystems, and snapshotting LVM 
partitions to simplify backup) I was running it in JBOD mode, and not 
using the motherboard RAID.  The LVM partition was the planned home of a 
large set of VMware VMs.

I am now running the machine as a Windows XP MCE box with no issues, and 
have a striped drive using XP's dynamic disk technology across three of 
the four drives in place above.  Tests so far are regular home user 
sorts of things, plus writing and then verifying a 55Gb backup using 
Windows backup (drive 1->2-4, then verifying). 

My conclusion, having nothing else to point to, is some subtle 
incompatibility between FC6 and SATA as implemented on ASUS motherboards.

Alan

-- 
Alan Peery                | Wisdom consists of knowing when to avoid
Systems Integrator        | perfection.
peery @io.com              |
07812 NNN NNN UK          | -- Horowitz




More information about the fedora-list mailing list