Re: Bad block repair
From: Thomas Schulz (schulz_at_adi.com)
Date: 12/21/04
- Next message: Erik Larsson: "Re: Xorg configuration in s10_69?"
- Previous message: Toomas Soome: "Re: "Torn between two OS" - Solaris vs Linux"
- In reply to: Andrew Gabriel: "Re: Bad block repair"
- Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]
Date: Tue, 21 Dec 2004 16:31:33 +0000 (UTC)
In article <cq9egi$i1e$2@new-usenet.uk.sun.com>,
Andrew Gabriel <andrew@cucumber.demon.co.uk> wrote:
>In article <cq7dpa$7oj$1@bluegill.adi.com>,
> schulz@adi.com (Thomas Schulz) writes:
>> We have an Ultra 10 with a Maxtor 6L080J4 connected to the internal ide
>> controller. The disk recently developed a bad block and I thought to
>> repair it using format. As the disk is only a few months old and there
>> is only one bad block, I thought that it would be safe to keep using
>> the disk. I used the analyze command in format to find the bad block
>> and to hopefully repair it. The block was found, but no repair was done.
>> Since the exact block was displayed in /var/adm/messages, I thought to
>> try the repair command in format. I got the following output:
>>
>> format> repair
>> Controller does not support repairing.
>> or disk supports automatic defect management.
>> format>
>>
>> The lines in /var/adm/messages are:
>>
>> Dec 19 12:18:02 seahorse dada: [ID 107833 kern.warning] WARNING: /pci@1f,0/pci@1,1/ide@3/dad@0,0 (dad0):
>> Dec 19 12:18:02 seahorse Uncorrectable data Error: Block 1abc99
>> Dec 19 12:18:04 seahorse dada: [ID 107833 kern.warning] WARNING: /pci@1f,0/pci@1,1/ide@3/dad@0,0 (dad0):
>> Dec 19 12:18:04 seahorse disk not responding to selection
>> Dec 19 12:18:04 seahorse dada: [ID 107833 kern.notice] dad0: disk okay
>>
>> Is this a limitation of this particular disk drive, or a limitation of
>> the controller in the Ultra 10? I had to change the partitions to avoid
>> the bad block and was unable to save the file system that contained the
>> block. Fortunately I had backups. Is there anything that I could have
>> done to repair this bad block and save the file system?
>
>IDE controllers normally remap bad blocks themselves
>(which is why format(1M) won't do it for you).
>It may actually have already done that, but until you
>write some new data to the block, it must keep returning
>the bad block status. Otherwise you would be under the
>false impression it had actually managed to read the
>original data from the block and potentially suffer a
>silent data corruption when you used it.
That was it. As soon as I used write instead of read as the command to
format analyze, the bad block disappeared.
>--
>Andrew Gabriel
>Consultant Software Engineer
-- Tom Schulz schulz@adi.com
- Next message: Erik Larsson: "Re: Xorg configuration in s10_69?"
- Previous message: Toomas Soome: "Re: "Torn between two OS" - Solaris vs Linux"
- In reply to: Andrew Gabriel: "Re: Bad block repair"
- Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]
Relevant Pages
|
