I agree with EZ, on a production system, I wouldn't use JBOD to hold data, nor back it up, hard disks these days are not that expensive, what more, on the suggested deployment, should a disk fail you will loose the content of both disks, whereas using any sort of redundant array, you will usually have a chance to recover most of the data.

Liran Cohen
TheOg

Ez-Aton wrote:
I would never use raid0 on a production system. Just not worth it.

However, both home system and raid1 are another thing.

BTW - You cannot boot from a raid0 partition. You need to have your /boot on a non-striping raid (that is, none, or raid1).

Ez./

Ira Abramov wrote:

Quoting Ez-Aton, from the post of Mon, 20 Dec:



as raid1. I can have swap partitions without RAID and the backup
directory as RAID0 instead of RAID1 for instance.

questions -
1. Is that the best methodology to follow? any other recommendations?



I would have recommended using raid0, but once upon a time, when a disk crushed, and I lost 80 gigs of data (these were the days when 40 gigs were large disks), I saw, or better - felt the negative results.



that's why I use that for the backup.

I stage a backup of the site on the RAID0 space, and then RSYNC it away
from there. plus this is a temp thig. I'm pretty sure we'll use a third
drive for swap and backup one day.



2. Am I risking anything by installing reiser3 (kernel 2.6.8) on an


MD device?



Should be no reason.



thanks, that's the part that had me worried the most.




================================================================= To unsubscribe, send mail to [EMAIL PROTECTED] with the word "unsubscribe" in the message body, e.g., run the command echo unsubscribe | mail [EMAIL PROTECTED]



Reply via email to