Home
| Calendar
| Mail Lists
| List Archives
| Desktop SIG
| Hardware Hacking SIG
Wiki | Flickr | PicasaWeb | Video | Maps & Directions | Installfests | Keysignings Linux Cafe | Meeting Notes | Linux Links | Bling | About BLU |
On Wed, Oct 02, 2002 at 07:12:32PM -0400, Andy Davidoff wrote: > Yesterday I built a SCSI RedHat-7.2 box with the root filesystem on > RAID1. Today I successfully booted from both disks without changing > a thing. If you are having problems with a boot-RAID, I'd look to > the fact that it's IDE or that the array wasn't configured during the > installation process. > > I have also lost disks under SCSI RAID1 without hanging the system or > causing unscheduled downtime of any sort. Gone are the days when to > solve RAID problems I had to hack kernel source. The stuff just works. Last night I was doing more experiments on this. I can unplug /dev/hda and get /dev/hdc to boot. The magic there seems to be to have a raid-extra-boot line in the file. Now I want to restore the contents of /dev/hda, so I plug in that disk, but I can't get /dev/hdc (which has good contents) to boot when there is a /dev/hda present. I am assuming it has something to do with what disks the bios reports as present and some "bios=0x81" variation will help. Actually, I can get the kernel to boot. I even see lots of cool messages fly by talking about how md's are starting up, but when the kernel tries to fire up init, it can't find it. It suggests giving a init= parameter to the boot, but I don't know what I might supply. It is already looking on the right md, I think. Another clue, when the boot works properly my "vga=0x303" works to give me small type, but when it is going to fail I get big 25x80 characters. Thanks for any suggestions, -kb, the Kent who figures he needs to read more about how PC bios's find and report disks.
BLU is a member of BostonUserGroups | |
We also thank MIT for the use of their facilities. |