[clug] Managing NAS units?

Robert Edwards bob at cs.anu.edu.au
Fri Nov 2 01:25:06 GMT 2007


Various quotes about the use of indirection to solve problems abound:
"Computer science is the discipline that believes all problems can be
solved with one more layer of indirection." - Dennis DeBruler

Anyway, your (Chris) proposed solution seems to add extra layers of
indirection (LVM being one of them) which will become both performance
bottlenecks and points of failure.

For my money, NAS/SANs are good for servers that need to share storage
at the filesystem or block level, but I don't see the need for
aggregating storage for separate services. Of course, the NAS vendors
will give you a different story.

A largish (unnamed) organisation lost it's NAS for a couple of days
earlier this year due to a mis-applied firmware upgrade and it took
out a number of significant but independent services.

My question is: why did those quite independent services need to
share the NAS? Why couldn't they each have a smaller NAS for primary
storage and possibly share a separate storage system for secondary
(backup) storage?

Cheers,

Bob Edwards.

Michael Still wrote:
> Chris Smart wrote:
>> -----BEGIN PGP SIGNED MESSAGE-----
>> Hash: SHA1
>>
>> Hi,
>>
>> Wondering if people have advice on the best way to manage a bunch of NAS
>> units on a network? I was thinking of using LVM to present them as a
>> single device and share from a server.
>>
>> Is this a good idea? Am I stupid? Ok, I'll re-phrase that..
> 
> That's pretty horrible isn't it? The intermediary server ends up having 
> to receive every packet for a file operation, and then send it out to 
> the actual NAS device.
> 
> Netapp does this with a crazy symlink tree thing. That's lame though.
> 
> Mikal



More information about the linux mailing list