[SATLUG] NAS HD recommendation

David Kowis dkowis at shlrm.org
Wed May 22 16:54:42 CDT 2013


On 05/22/2013 04:31 PM, Alex Bartonek wrote:
> 
> 
> --- On Tue, 5/21/13, Bruce Dubbs <bruce.dubbs at gmail.com> wrote:
> 
>> From: Bruce Dubbs <bruce.dubbs at gmail.com> Subject: Re: [SATLUG]
>> NAS HD recommendation To: "The San Antonio Linux User's Group
>> Mailing List" <satlug at satlug.org> Date: Tuesday, May 21, 2013,
>> 10:00 PM
> 
>> I think spending extra for 'industrial' drives is not a 
>> particularly good expense.  That's what you have backups/RAID.
> 
> 
> I was thinking just running regular SATA HD's which would save me 
> $50/drive.  I hear good stories and bad stories about drives, I
> guess I wanted to be safe.  But you're right..I'll just make sure I
> have enough spares and just monitor SMART data that my morning
> reports have on my server.  If you're thinking "morning
> reports??"..LOL.. I have 2 servers at home running 24/7 and yes, I
> have them generating reports for various things like SMART data,
> dmesg output, disk space etc and I look at them every day.

http://storagemojo.com/2007/02/19/googles-disk-failure-experience/

Might be interesting to note that SMART data doesn't actually start to
indicate failure. It might tell you when it's already failed, but it's
not likely to tell you that it's about to fail.

ZFS provides some measure of catching about drives that are going to
fail, because it catches bitwise errors thanks to it's resilvering
logic. At least, it's the only RAID mechanism that I know of that will
catch those kinds of errors.

If I had time to reset everything, I'd do my raid array on ZFS, using
RAID-z and let it keep an eye on my drives and their data for me.

I might even do this anyway, since zfsonlinux is good enough to use
for safe data. Maybe not really fast, certainly not as fast as the
Illumos implementation, but I need Xen, and that's on linux, so there
you go.

--
David

> 
> -Alex
> 
> 



More information about the SATLUG mailing list