I am only using a partition of the disk (sda3, sdb3) for the raid,
so I don't want to reinstall everything. Any other way to wipe out
the persistent "superblocks"? I tried to "mke2fs" on the patition, which did
not work.
When I tested "raidsetfaulty, raidhotremove, raidhotadd" on raid1,
I also got problems. I did it successfully a few times, but most of the
times the raid behaved abnormally. (recorey stay at 0%, read/write
on the raid was blocked...)
(I can give more detail infor, it is *easy* to reproduce.)
It seems that linux raid is not stable enough for any production
environment. A little disappointed about that. (How about windows
NT?)
Dong
At 08:18 PM 11/25/99 -0500, you wrote:
>This (almost same) problem happened to me as well. It was solved with a
>wipe of all information from the disk by using DOS fdisk to eliminate the
>partitions and overwrite the MBR with fdisk /mbr . That worked and I was
able
>to create the new raid configurations.
>
>Could it be that the "persistent superblocks" are a bit too persistent?
>
>alex
>
>
>
>On Thu, 25 Nov 1999, Dong Hu wrote:
>> I am testing raid tools on linux and I have a problem.
>>
>> I have two same SCSI harddisk and have one same size
>> partition on each harddisk, sda3 and sdb3.
>>
>> I configured sda3 and sdb3 to a raid 1(mirror) with no problem.
>>
>> Now I want to change the configuration to raid0,
>> so I edit the /etc/raidtab file,
>> issue mkraid --force /dev/md0,
>> everything seems fine, but when "raidstart" and check the status,
>> the raid is still running in old raid1.
>>
>> I just could not change the configuration to anything other than
>> raid 1.
>>
>> kernel: 2.2.11 with patch raid0145-19990824-2.2.11.gz,
>> raidtools: raidtools-19990824-0.90
>>
>> Any idea?
>>
>> Thanks,
>> Dong Hu
>
>
