On Mon, 22 Dec 1997, Eric S. Hvozda wrote:

> If you have a second hard disk, this would seem to be a simple task:
> 
> *  Create a filesystem large enough to hold the current / on the second
>    hard disk
> *  Newfs, fsck the new file system
> *  Use dump/restore to copy the existing / to the new file system
> *  Reboot single-user to the new / (on the 2nd disk)
> *  fsck and mount / and /usr
> *  Resize / on the 1st disk to the desired size, newfs and fsck
>    (swap will probably shrink as a result)
> *  Use dump/restore to copy / back to the 1st disk
> *  Reinstall boot blocks (this is very important)
> *  Reboot, as per normal operation; / is now resized.

This is what I was already thinking, but "simple" is not the adjective I 
would use to describe this! 

I have a second disk, but it has a FAT16 filesystem on it for Win95/DOS. 
Right now I am thinking I should just backup my files to the second 
disk and  wait for 1.3 to be released. That way I can experiment with the 
new install procedures on my primary disk. Once I have a 1.3 system up, I can 
then selectively restore the files I need.

> > IBM's concepts of a volume groups, logical volumes, etc..., suddenly seem 
> > to make a lot of sense...
> 
> Well yes, and no.  It makes sense until you want to make / live on
> a logical volume.  What are you gonna do if for some reason you
> can't get /'s volume group to activate and you don't have an

Well, in the NetBSD-world, if your / file system was trashed and wouldn't
boot, you would have to restore from a backup. In the same way, with 
AIX's volume groups, if your root volume group won't activate because it 
is trashed, you would restore from a backup. However, one thing I like 
about the backup for an AIX root volume group is that you can make it 
bootable. Problems with the root volume group? No problem. Shutdown, put 
the computer in service mode, put your backup in the tape drive, reboot, and 
follow the prompts. At work, I manage about 40 RS/6000 workstations and have 
had to do this a couple of times due to disk drive failures.

> We currently have the ability to have disks concatenated/stripped/mirrored
> together via ccd(4).  However, the complexity required to allow things

This is something that I need to look at again. When 1.2 was released I 
remember seeing a lot of messages about problems using ccd, so I stayed 
away from it. I believe there have been a number of fixes in this area, 
so hopefully it is much more stable.


Brian

-------------------------------------------------------------------------
| Brian Stark                       | Internet : bstark@siemens-psc.com |
| Siemens Energy & Automation, Inc. |          : bstark@empros.com      |
| Power Transmission & Distribution |                                   |
| 7225 Northland Drive              | Voice    : +1 612 536-4697        |
| Brooklyn Park, MN 55428   USA     | Fax      : +1 612 536-4919        |
-------------------------------------------------------------------------