[wellylug] Testing hard disks

[vincent cox]

Hi

I have had these errors in the past.  If I remember correctly I was 
using the wrong ide cable for the drive.

The linux system I was using at the time worked ok but /var/log/messages 
was full of these errors.

Just a thought.

vince

>>- when there's some sort of DMA issue.
>>    
>>
>
>Hm. Yes all the errors are DMA errors:
>
>Sep 16 09:58:10 Rescue kernel: hda: dma_intr: status=0x51 { DriveReady SeekComplete Error }
>Sep 16 09:58:10 Rescue kernel: hda: dma_intr: error=0x40 { UncorrectableError }, LBAsect=1029694, sector=65704
>Sep 16 09:58:10 Rescue kernel: end_request: I/O error, dev 03:02 (hda), sector 65704
>
>Sep 16 10:00:30 Rescue kernel: hda: DMA disabled
>
>Sep 16 10:00:35 Rescue kernel: hda: read_intr: status=0x59 { DriveReady SeekComplete DataRequest Error }
>Sep 16 10:00:35 Rescue kernel: hda: read_intr: error=0x40 { UncorrectableError }, LBAsect=1029696, sector=65796
>Sep 16 10:00:35 Rescue kernel: end_request: I/O error, dev 03:02 (hda), sector 65796
>
>But turning DMA off makes no difference. And the whole box has run
>flawlessly for months without trouble, with DMA on. A hard disk without
>DMA is not that useful. Besides, DMA works fine on the other 9G
>partition.
>
>  
>
>>going on, but I'd expect that badblocks -w, which simply writes to each
>>block, flushes the cache, and reads it back in again (4 times, with 4
>>different patterns) would trigger the same issues as an I/O install,
>>since it's going through the same hardware calls.
>>    
>>
>
>That's what I would have expected too. Fact is, badblocks -w finds nothing
>wrong, neither does reading the whole partition with dd, after having it
>nulled out. Before nulling it out reading it with dd would throw one
>error after the other, with the errors being clumped together in some
>range, and other block ranges being error free.
>
>Volker
>
>  
>