Hi Jggimi. Your post is more than I could have hoped for, thanks!
Quote:
Originally Posted by jggimi
This holds true with encrypted data. The same three causes for data loss -- human, hardware, and software -- apply. There may be more opportunities for human error, as you're adding a configuration layer. There is a minor opportunity for software error, though
|
Makes perfect sense! I'm going for a full disk encryption to keep it simple and avoid the most possible human errors. If I share my encrypted data with the root (/) directory I can't just plug it out and move it to another box for example.
Quote:
Originally Posted by jggimi
Over the years since its incept, softraid(4) has had changes to stored metadata which , when we upgraded, required us to backup - recreate - restore softraid entities. I expect that to continue, as development is not complete for all disciplines. |
I especially appreciate this piece of advice! I was going to backup anyway but this might change it to a certain extent!
Quote:
Originally Posted by jggimi
I consider the crypto discipline fairly mature, and I've been using it in production for at least four years. It was the second discipline added to softraid, in May 2007.
|
Good seems to be more than stable enough for what I need. The book gave me the impression that the encryption was more less experimental.
Quote:
Originally Posted by jggimi
If you have the resources, you can use the crypto discipline atop an array created with the RAID1 discipline.
|
Might be a good thing to think about. I've read the thread you linked to. Seems like human error and configuration are the most vulnerable things indeed. I'll see what I'm most comfortable with, a backup solution or RAID one with a backup solution.