[sclug] Re: sclug digest, Vol 1 #162 - 14 msgs

Neil Haughton n.a.haughton at bigfoot.com
Sat Oct 25 09:05:45 UTC 2003


>
> Subject:
> Re: [sclug] RAID 1 adventure
> From:
> lug at assursys.co.uk
> Date:
> Sun, 22 Jun 2003 19:23:34 +0100 (BST)
> To:
> Neil Haughton <n.a.haughton at bigfoot.com>
>
>
>On Sun, 22 Jun 2003, Neil Haughton wrote:
>
>  
>
>>I have my Windows drive C: on /dev/hda1, but it expects to find drive D: 
>>what was my /dev/hdc1 which is of course no longer there. (C: is OS and 
>>apps, D: contains all the data files). How do I tell Win98 to look on 
>>partition /dev/hda10 for its drive D:?
>>    
>>
>
>I don't think you can. Windows (9x and below) names the drives by iterating
>through all partitions it recognises over all drives present:
>  
>

Actually, to my surprise it worked pretty seamlessly. I simply started 
Win98 with the new partitions in place and copied over and the old sec. 
drive disconnected, and it saw the 2nd FAT32 partition as D: without a 
murmer (simply ignored the existence of the in between Linux 
partitions).  So, not a completely bad OS then :-)

>  
>
>>....he idea is that after Windows is 
>>used, Linux will resync the mirror drive when it nexts starts up, as 
>>long as the Windows partitions are mounted. I trust I have this right?
>>    
>>
>
>I haven't tried that usage, but I _suspect_ it won't work as the partition
>type will be 0xfd rather than one of the FAT types. OTOH, I'm not sure
>Windows will care as long as it finds something that looks like a FAT
>filesystem inside... :-]
>
>Also, I'm not sure of any way to specify to the RAID layer which drive is to
>be the "canonical" copy. You might find that on the first boot it copies
>the contents of your empty partition over the working one!
>

I thought that the raidtab file defined this,

    device      /dev/hda1
    raid-disk   0
    device      /dev/hdc1
    raid-disk   1

Doesn't this mean that raid-disk 0 is the canonical disk?

Regards,

Neil Haughton.





More information about the Sclug mailing list