"snapshot" RAID

[tmetro-blu-5a1Jt6qxUNc at public.gmane.org (Tom Metro)]

Dan Ritter wrote:
> Tom Metro wrote:
>> How about sharing the code you use to implement your backups?
> 
> I didn't write any code for this -- well, mostly. The rsnapshot
> cron job calls ~dsr/bin/rsnapshot.wrapper, which runs rsnapshot
> and then immediately runs hdparm to spin down the disk again.

That's all I meant. The wrapper script. Perhaps the hdparm commands are 
obvious (I haven't checked), but I figured it'd be good for anyone 
interested in this to see a full example with hdparm and rsnapshot.


>>> (sdb1 is a copy of the boot partition, tested and then never mounted.)
>> But that needs to be re-mirrored after each kernel update, no?
> 
> No. I just need it to be able to boot enough to access the disk
> and network. At that point I can figure out why I couldn't boot
> from the first disk, replace it and reinstall the OS, if needed.

Ah, so you're just using the second disk as 1. a bootable rescue disk, 
plus 2. snapshot archival storage.

I'd rather have a second disk be as close to a mirror as possible, so if 
something fails on the primary, you can immediately continue productive 
work with the secondary (after a reboot, of course). (The secondary 
would still serve the purpose of a rescue disk, and you can use it to 
investigate the problem later.)

With UUID being the standard way to find partitions these days, you'd 
have to use a different mechanism (adjust GRUB config, fstab), or tweak 
the second drive after cloning the primary. But with bootable RAID 
mirrors being somewhat common these days, I'm sure this is well traveled 
ground.

  -Tom

-- 
Tom Metro
Venture Logic, Newton, MA, USA
"Enterprise solutions through open source."
Professional Profile: http://tmetro.venturelogic.com/