Boston Linux & Unix (BLU) Home | Calendar | Mail Lists | List Archives | Desktop SIG | Hardware Hacking SIG
Wiki | Flickr | PicasaWeb | Video | Maps & Directions | Installfests | Keysignings
Linux Cafe | Meeting Notes | Blog | Linux Links | Bling | About BLU

BLU Discuss list archive


[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

NAS devices



On 06/01/2010 08:11 PM, Edward Ned Harvey wrote:
> Phew ... it's really unfortunate that you have a preference for Linux as a
> NAS.  Because it's certainly not the best solution, unless you have some
> really specific linux-only application, or something like that.
>
> In terms of performance, reliability, speed, backups, and every
> characteristic that I can basically think of mattering in a NAS, I would say
> solaris/opensolaris/ZFS would be the better solution.
>
> I'll only go into it further upon request (maybe I'm already wasting my
> breath) but I'll highlight the reasons I think are the most important here:
>
> * ZFS supports snapshots.  This means, 99% of the time a user needs or wants
> to have something restored from backup, they're able to simply copy it from
> another directory.  They don't need any assistance.  It's already available.
>
> * Also because of snapshots, incremental backups and "live" backup servers
> are infinitely faster and more possible than linux.  I formerly used rsync
> on a RHEL box, spending 10 hrs nightly just to walk the tree and scan for
> changes to send to the backup server each night.  Now, my nightly
> incrementals take an average 7 mins, which varies based on quantity of
> changed data that day, because there is no necessity to walk the tree
> scanning for changes.  ZFS already knows which blocks changed, and is able
> to simply stream one large sequential data stream.
>
> * ZFS is able to do software raid faster than the fastest hardware raid
> controller.  This is because hardware raid solutions only know about device
> blocks; they don't have knowledge of the filesystem.  In ZFS, the OS has
> knowledge of both the block devices and the filesystem, so ZFS is able to
> aggregate many unrelated small block writes into a single large sequential
> block write.
>
> And that's good enough.  If this hasn't sold you on ZFS, it's a lost cause.
> ;-)
>
>
>    

I'll look into ZFS. Our product runs on Solaris as well as Linux and 
AIX. The issue here is budget and politics. We don't want corporate IT 
involved in our operations, but we might be able to get an adequate 
Solaris box from Toronto. I think that Toronto may be converting to SAN, 
and in that case, maybe one of their current systems could be grabbed.

-- 
Jerry Feldman<gaf-mNDKBlG2WHs at public.gmane.org>
Boston Linux and Unix
PGP key id: 537C5846
PGP Key fingerprint: 3D1B 8377 A3C0 A5F2 ECBB  CA3B 4607 4319 537C 5846







BLU is a member of BostonUserGroups
BLU is a member of BostonUserGroups
We also thank MIT for the use of their facilities.

Valid HTML 4.01! Valid CSS!



Boston Linux & Unix / webmaster@blu.org