Re: 10.0 - Installer newfs segfaults on SS2

[vom513 <vom513%gmail.com@localhost>]

> On Sep 1, 2024, at 5:14 PM, Mouse <mouse%Rodents-Montreal.ORG@localhost> wrote:
> 
>> Disk /dev/loop19: 8.43 GiB, 9056690176 bytes, 17688848 sectors
> 
>> /dev/loop19p1        0 17619361 17619362  8.4G  0 Unassigned  
>> /dev/loop19p2 17623305 17688840    65536   32M  0 Unassigned  
>> /dev/loop19p3        0 17688847 17688848  8.4G  0 Unassigned    u  
> 
> You said upthread that
> 
>> The exact size of the 8G empty image I use on the ZuluSCSI is
>> 8589934592.
> 
> This now reports 9056690176 bytes (according to both the header line
> and the largest size figure).  Which, if either, is actually correct?
> 
> If you're labeling a 8589934592-octet disk with a label causing
> userland to expect 9056690176 bytes, this will cause _some_ kind of
> trouble.  I wouldn't expect that trouble to be a coredump, just write
> errors, but perhaps there's a bug somewhere in the error path.

You are correct.  Somehow the image on the SS2 is “growing” ?

This is my “source” blank image:

vom@ocl:~$ ls -la Downloads/HD30_512-8gb-nbsd-BLANK.hda 
-rw-r--r-- 1 vom vom 8589934592 Apr  5 17:25 Downloads/HD30_512-8gb-nbsd-BLANK.hda

This is the SS2’s image:

vom@ocl:~$ ls -la /media/vom/SS2-NBSD/HD30_512-blank.hda  -rwxr-xr-x 1 vom vom 9056690176 Aug 30 11:32 /media/vom/SS2-NBSD/HD30_512-blank.hda

This is from an SS10, ZuluSCSI happily running 10.0.  No “growth”:

vom@ocl:~$ ls -al /media/vom/ZS-TEST/HD30_512-8gb-nbsd-BLANK.hda  -rwxr-xr-x 1 vom vom 8589934592 Aug 25 13:26 /media/vom/ZS-TEST/HD30_512-8gb-nbsd-BLANK.hda

What on earth could be causing this ?  The fact that a real Seagate HDD had the same behavior makes me think it’s the SS2 and not the ZuluSCSI.

I guess I need to “setenv random-disk-growth? false” ? :P

Thanks to everyone so far for the replies, insights and patience.