[EMAIL PROTECTED] wrote: > You might look into the kernel software RAID if you're running kernel > 2.4.x. It supports RAID-1, which is mirroring. Although mounting a
Yes, I'm running on 2.4.x. > disk and doing a manual copy would work, in the event your system disk > fails you'd be stuck with an unbootable system. With a RAID, you can > failover to the second disk and continue running while you work on > replacing the failed disk, then resync everything automatically. This sounds good. Then what tools can I use to set it up? (Assuming that kernel support for RAID-1 is done). BTW, the controller is SCSI, not RAID; I don't think that I'd be able to hot-swap the disk. This sounds really interesting; failover just like that...? And then resync...? Wow. But the above "fail" means the disk is no longer usable right? What about the /dev/sd<x> thing? If the first one fail, would the second still be /dev/sdb? My concern is what's in /etc/fstab. Besides, that failover sounds like going to work for data disks. I'm talking about system disks here. What I have done is to create a single partition on the second disk (my swap is on a file), copy all the / files into it (using cp -avf), and then create a new entry in the Grub's menu.lst (which resides on the first disk); hoping that there would be no (hardware) disk failure. I think this would cover me on events like filesystem failures. I use ReiserFS, BTW. It's _way_ fast, but even in 2.4.x it's not considered as stable. The machine has been up for about 7 days; the last reboot was due to out of swapspace, I believe. Or some holes on the swap partition (I have put it on a file, recently). Another thing, is there any way to chroot apt-get? I'd like to install new packages, on the second disk, but using /mnt as the root directory. TIA, Oki