View Single Post
Old 31st January 2012
gillindu gillindu is offline
Port Guard
 
Join Date: Jan 2012
Posts: 36
Default

Quote:
Originally Posted by IdOp View Post
When you boot the "real NetBSD" (on disk), are the outputs of fdisk and disklabel what you think they should be? In other words, does it appear that the disklabel update from your protofile is being used by the rebooted system?
Disklabel, as far as I can see (maybe I'm missing something) - yes. Moreover, to make room for one of these partitions in disklabel, I've pulled off the ex-wd0j and made the ex-wd0i - new wd0j (just to have a kind of marker). And yes - /dev/wd0j can be mounted and it is what is shold be. So, the new disklabel table has been accepted.
But, fdisk... as the first thing it says that "Extended partition table is corrupt". And then... I'm still trying to understand something (it's a rather long output). The sizes seem to be as they should be, but, one of them seems to start nearly where it should (1 or 2 sectors of shift, I'm not sure), while the another one seems to be completely misplaced! And now, looking at the sfdisk output (from Linux), it seems to be... my mistake! I've copied from the wrong columns.
So, now, when I try to run "disklabel -R wd0 newproto" (with the corrected values)... I'm getting again that wd0g and wd0h overlap! (Now it would be for 62 sectors and not 126 as before.) After the sfdisk output, there should be 28 sectors of free space. It seems again that the size of the NetBSD's slice (wd0c) is slightly mispresented in the partition table, so the Linux utilities get fooled (at least, that would be my understanding)
(And that's why I didn't go directly to expand the slice, but, I've tried first with this "placeholder"! Better that such things happen now!)
Reply With Quote