Subject: Re: 36GB / , help!
To: Piotr Stolc <socrtp@soclab.eu.org>
From: Manuel Bouyer <bouyer@antioche.lip6.fr>
List: netbsd-users
Date: 12/22/2003 19:18:19
On Mon, Dec 22, 2003 at 01:28:46PM +0100, Piotr Stolc wrote:
> On Sun, Dec 21, 2003 at 09:38:39PM +0100, Manuel Bouyer wrote:
> > > is there any way to preserve first sectors on / to install new kernel in
> > > the future? or maybe there is a method to shrink rw-mounted fs?
> > 
> > Unless you machine is really old, you should not have any problems loading
> > the kernel from the end of disk. The NetBSD boot loader has been able to use
> > LBA when available for a long time.
> 
> yes, I remember problems with booting on P1 and P2 machines with IDE drives.
> but this is P3 1.4 GHz machine with SCSI RAID controller:
> 
> mly0 at pci2 dev 8 function 1: Mylex AcceleRAID 170
> mly0: interrupting at irq 9
> mly0: 1 physical channel, firmware 6.00-7-00 (20001214), 32MB RAM
> scsibus0 at mly0 channel 0: 16 targets, 1 luns per target
> scsibus1 at mly0 channel 1: 16 targets, 1 luns per target
> scsibus2 at mly0 channel 2: 16 targets, 1 luns per target
> [...]
> sd0 at scsibus1 target 0 lun 0: <MYLEX, RAID 1, ONLN> SCSI3 0/direct fixed
> sd0: 34984 MB, 4459 cyl, 255 head, 63 sec, 512 bytes/sect x 71647232 sectors
> sd0: sync (50.0ns offset 8), 16-bit (40.000MB/s) transfers, tagged queueing
> 
> would NetBSD boot loader have any problem loading kernel from the end of
> disk on this configuration?

I expect the controller to support LBA ...

> that's why I always create a small / partition. but this machine is already
> installed and the biggest problem is that I don't have physical access to it
> - I can manage it only through SSH. but if this controller is good and it
> will boot kernel from any sectors correctly, then I'll just leave it as it
> is now.

Just look at the block numbers of /netbsd with fsdb. If you already have
blocks above 8G, you're already using LBA properly do boot :)

-- 
Manuel Bouyer <bouyer@antioche.eu.org>
     NetBSD: 23 ans d'experience feront toujours la difference
--