Re: RAIDframe corruption

[Emmanuel Dreyfus <manu%netbsd.org@localhost>]

On Tue, Mar 01, 2016 at 07:07:58PM +0700, Robert Elz wrote:
>   | - 64 MB region at offset 498074652673 with many changes

> You have 500GB dives, right?  So that is way out near the end.
> What number(s) of sectors do the drives report (should be in dmesg) ?

Yes, 500 GB disks, 1953525168 sectors. The offset above is around 463 GB.

> What is in the labels for wd0 and wd1 (MBR probably, and the netbsd disklabel
> if there is one, but GPT would be OK too - whatever you are using).
> Note I mean the labels on the raw drives, not what is inside the raid
> "device", that is irrelevant for this.

Oh yes, you are right. I am looking at difference between swap
partitions:

#        size    offset     fstype
 a: 972800000      2048       RAID
 b: 980723120 972802048       swap
 c: 1953523120      2048     unused
 d: 1953525168         0     unused

Hence, the difference in the RAID partitions are:
- 32 bits at the end of MBR bootstrap 
- one bit in RAIDframe structures

But I still have no explanation why the kernel got corrupted and if that
problem could be more widespread. RAIDframe is probably innocent there,
though.

-- 
Emmanuel Dreyfus
manu%netbsd.org@localhost