[Hcoop-discuss] Server pricing

[Justin S. Leitgeb]

The problem is that if we're really talking about a terabyte of storage 
space or more, I think that it is still outside of our budget to buy 
space that is both abundant *and* fast.  If you don't believe me, try 
pricing out a system with 1 TB of SCSI storage in RAID 10, a powerful 
RAID controller, and a 2U+ machine to house it in.  By the time we came 
up with $6,000+ US to spend on such a system, we would have outgrown it 
in terms of our space needs and have to buy another bigger and faster one. 

The only way that I can imagine buying a large storage system is to pack 
a box with SATA disks, which is going to be less reliable and slower 
than a SCSI equivalent.  That would make it unreliable as a login 
server, and more useful as an "offline" storage box for hcoop members 
(i.e., space not actively being hit, aside from occasional transfers).  
While I think that this could be useful in the future, it would be fine 
to add to our services later, rather than putting it in a central place 
in our environment.

Personally, I would still rather get a strong web server, a strong IMAP 
+ hcoop services server, and then think about our next upgrade, which 
would probably be a hardware firewall, or another web server rather than 
a storage server.

This said, I'm not against storage servers in enterprise environments... 
but a SAN is a huge investment, in time and money, and we're not ready 
for it yet.  A NAS (basically a dedicated fileserver box) is more within 
our means, but as a later add-on, rather than a central part of our 
environment.  This is basically because we are not yet ready to lay out 
massive amounts of cash for space that is abundant and fast.

I know that one of the reasons that we were looking at an architecture 
that gave a fileserver a central place was that it would make 
administration easier... but we have to recognize all of the tools that 
are available to us, as unix/linux admins in a growing network.  I'm 
thinking especially of tools like cfengine, which I'm deploying now to 
help manage a network of 300 + linux boxes and a handful of Sun machines.

Sorry if I'm being problematic by not subscribing to the fileserver idea 
easily, I just don't see the benefit of it right off the bat as a 
central piece of our infrastructure.  Let's grow a bit first, and then 
wait until we actually have usage patterns that indicate where we could 
leverage it, and then start looking at options.

If we do want the terabytes of storage now, the only option that I see 
is a SATA RAID array, that is set up like a NAS -- still not in the 
center of our network.  This is the only tenable option that I see right 
now.


Justin


Adam Chlipala wrote:

>Justin S. Leitgeb wrote:
>
>  
>
>>Based on our discussions, and from looking at our utilization of fyodor, 
>>buying two of these (one for login and user files, the other hcoop 
>>administration and services, including IMAP) should keep us going for a 
>>while quite comfortably.  So, a starting hardware package, in my 
>>estimate, would cost $4224, with basic redundancy and room to grow.
>>
>>    
>>
>So you don't think it's appropriate to have a dedicated file server?  I 
>thought that we had definitely agreed on using a shared filesystem in a 
>previous thread.  Are you proposing that the "hcoop administration and 
>services" server be the one with the terabytes of storage attached to 
>it?  (I have no doubt that even the current membership size could 
>gainfully use a terabyte of disk space if we announced that we could 
>support arbitrary amounts of it.)
>
>_______________________________________________
>Hcoop-discuss mailing list
>Hcoop-discuss at hcoop.net
>http://hcoop.net/cgi-bin/mailman/listinfo/hcoop-discuss
>  
>