On Thu, Nov 17, 2011 at 11:47 AM, Neil Bothwick <n...@digimed.co.uk> wrote: > On Thu, 17 Nov 2011 10:59:06 -0800, Mark Knecht wrote: > >> Yes, resize would be done to /dev/md?. I agree. However I don't >> believe that I'd use Neil's suggestion of fdisk block numbers on >> /dev/md, right? That doesn't make sense to me and I don't beleieve >> Neil was suggesting anything like that. > > Yes I was. /dev/md? is still a block device, and its blocks correspond to > physical blocks on the component drives. > > > -- > Neil Bothwick
OK, so returning to your original response, you suggest increasing the size of each physical partition and then resizing each of the physical partitions independently? (/dev/sdwhatever instead of /dev/md6 directly?) Is there a reason or personal experience you have to not to resize the RAID-5 directly? I completely trust you as to date I cannot remember anything you suggested I do that wasn't a good way to do it but doing /dev/sdwhatever seems problematic if it had been an 8-drive RAID-1 becoming a RAID-5, etc. - Mark <quote> Using cfdisk or fdisk, delete the partition and recreate it, USING THE SAME START BLOCK at a larger size. Then "resize2fs /dev/sdwhatever" will resize the filesystem to fill the partition. </quote>
