On Sat, Oct 30, 2010 at 12:18:42PM +0200, Jean-Francois wrote: > Le Saturday 30 October 2010 04:52:35, Marco Peereboom a icrit : > > On Thu, Oct 28, 2010 at 10:41:52PM +0200, Jean-Francois wrote: > > > > # bioctl -R sd0a sd2 > > > > > > If I understand well the above command kicks off a rebuild on a > > > replacement device. Few questions from my side ... > > > > > > Is it possible to rebuild with another device for example sd0b or sd1a > > > instead of sd0a ? (seems no if I understood properly) > > > > Assuming I got the question right, yes. You can rebuild on any > > appropriately sized chunk. > > > > > Is the same process as for initialization required for the rebuild ? e.g. > > > # fdisk -iy sd0 > > > # printf "a\n\n\n\nRAID\nw\nq\n\n" | disklabel -E sd0 > > > > > > Regards > > Ok, to be more clear, say for example we set a softraid called sd2 with > chuncks sd0a and sd1a in raid 1. > sd0a becomes faulty/offline, I would like to use an appropriate chunk to > relpace > it, but say it is not designed sd0a but sd3a, what can we then do ? Could we > rebuild on sd3a ?
Lets say you have a raid 1 made out of 3 chunks; sd1a, sd2a and sd3a. Now lets say sd2a breaks and you add a new drive sd4. On that new drive you create a d partition that is of the right size. You could rebuild the softraid volume with sd4d. I left out the drive shuffling that might (will) happen to simplify the example. To prevent shuffling from biting you in the butt read up on the DUID stuff that jsing wrote. It is described in the mount(8) page. > > Thanks & regards
