Hello World:

I am installing FC10 from live CD. 32 bit. 1 Gig of RAM. Semperon 3100+. Gigabyte mobo w/nforce 250 chipset and Nvidia 7600gs 8x AGP video card.

Don't laugh. If this works, it will be the most powerful node on my as yet to be assembled network.

So I have these two hard drives burning a hole in my pocket. Their identical and I want to use software raid however, I am not sure how to make sure these two fine drives are not going to waste.

On install, I click the raid button and go to "Add a Partion" for software raid and the default is 200 meg, and I am only allowed to use one drive, not both. Under additional size options I can set "Fixed size", "Fill all space up to (MB):" and "Fill to maximum allowable size".

So:

Why am I not going to be using both drives?
Why am I not going to punch in 320000 in stead of accepting the default of 200? It seems to me that what I think I should want to do is "Fill to maximum allowable size", Yes?/No?.
If I have FC10 on hda, do I want to select that drive for software raid also? or do I put the software raid on hdb which has nothing on it?

So, obviously I don't get it.

This is going to be a general purpose workstation just for me to learn with.

Should I (can I?) mirror the 320 Gig drives? That would be fine with me. 320 gig is more than enough so, why not just use the second drive for backup. That would certainly make sense if all I had to do was swap SATA cables and reboot in case of some catastrophic end user failure.

The other obvious choice would seem to be to combine them into "one". Does software raid combine both drives into one logical volume?

Ultimately I am looking for some ideas from the fedora peanut gallery as to what you would do with linux if you had two 320 gig drives sitting around but, with only the modest system that I have at my disposal.

tayl

The short qn is do you want high Availability, or Resiliency.
Availability = RAID 0 both drives
Resiliency = 1 active disk, 1 backup. You can probably get several generations of backup onto the 2nd disk using incremental+compression (rsync+gzip).
Also gives you the chance to take it off site / build a new box based on backup disk.

Note that you can still experiment with RAID by partitioning the first drive and RAID the partitions. Of course HW error would kill it, but that's what the backups are for, right?

I absolutely want to be able to take it off site and build a new box based on the backup disk.

So then, I would software raid everything? on hda and make sure that hdb gets mounted and back everything up onto hdb and, probably just do a full backup every time?

I am still perusing the online Red Hat tutorial ...

tayl