Slashdot Mirror

Step 1: Buy yourself something like this: http://www.aberdeeninc.com/abcatg/Stirling-X339.htm
Step 2: Install it
Step 3: rsync
Step 4: Go do something else -- this is going to take a while

When I google JBOD, from the *first page* of results:

http://www.aberdeeninc.com/abcatg/kitjbod-1003.htm

http://opensolaris.org/jive/thread.jspa?threadID=83108&tstart=0

http://blogs.sun.com/brendan/entry/jbod_analytics_example

Even for those running Windows, the Mac Pro is one of the most powerful computers available, bar none.
Depends on the workload I guess.

The mac pro maxes out at dual-quad (8 cores total) there are systems availible with quad-hex xeon processors but they are from the old core2 based generation ( https://iceberg.aberdeeninc.com/AberSys/ConPag.aspx?Nire=conpag&Tikya=scStirling444 ).

I can't find anything current gen with more than dual-quad so I guess which is better will depend on your workload.

Things get a lot more expensive when you go beyond two sockets though.

2 boxes for hardware failover will do you fine, if you are worried about HA the its the COST of downtime that you are worried about (i.e. down for an hour exceeds $1000 in lost revenue) which will justify the solution. Don't just drive availability to five nines because you feel its cool, do it because the business requires it.

This is something that is rampant: techies tend to overestimate the value of uptime.

Sure, it's sexy to have high availability this and redundant that, but unless your company is pulling down at least $1,000,000 per year or more in gross revenues, it's hard to beat the 3 to 4 nines or so uptime delivered by a good quality, whitebox server running Linux. Something like this unit would deliver excellent performance and excellent reliability at a very low cost.

How much does an hour of downtime actually cost your company? Be honest. If you had to tell your customers: "we were down for 2 hours because a software update caused us to have to ..." what would it actually cost your company? Especially if it only happened every year or so? In my experience, even in fairly stiff production environments, there has been no cost at all. We've maintained about 99.95% uptime for the past 3 years, with 1 "incident" every year or so, with no cost at all. In fact, our company has a good reputation for availability and support!

So don't spend money on sexy hardware with lots of blinkie lights and cross-connects, which often decrease your reliability by introducing unnecessary complexity.

Instead, spend money on your hosting. Don't *ever* host it in-house. Ever. Get a first-tier hosting facility, with redundant network feeds, power, and staff who give a damn. Don't be afraid to pay for it, because it will probably save you money, anyway. You'd be amazed at how price-competitive top-notch hosting farms can be!

Make sure to get to know the on-site techies on a first-name basis, give 'em a six-pack of their favorite beverage, and thank them profusely when they do anything for you. The goodwill these types of things can bring will work wonders for you down the road.

And remember:

2 nines is 3.65 days of downtime per year.
3 nines is .365 days of downtime per year (~ 8 hours)
4 nines is .0365 days of downtime per year (~ 45 minutes)

It's a very, very rare case indeed where 3-4 nines of uptime isn't completely sufficient.

And 1,000 unique visits per day? Pssht. Unless you are doing some pretty ferocious database stuff, (EG: joins across 12 tables with combined inner/outer/composite joins) the aforementioned server should do the job just wonderfully.

DON'T FORGET BACKUPS! And backup your backups, because backups fail, too.

About a month ago this blue screen appeared on a Windows 2003 Server Appliance Edition NAS BOX. Attempt to mount an NFS share served by the windows NAS, and boom blue screen. This is a commercial NAS box running a supposidly ultra stable version of windows for such devices. I would have picked the linux powered box but it wasn't my choice.

looks like if you look at the lower end itanium 2 (1.4GHz and less cache) you can get one for $900. the motherboard might cost $1k+ though.

here

Penguin Computing's still about.
Tatung seems to be making 1U Opteron rackmounts.
As does Aberdeen...
And Opteronics...
As is Aspen...

All of these vendors were found on the first page of a Google search of "1u opteron rackmount".

Of course, these aren't "major" players like HP, Dell, and IBM. Doesn't make the box any less reliable and you can buy support services from people like IBM, etc. for them anyhow.

Aberdeen sells their scalable NAS system that holds 3TB per 2U unit. It's expandable up to 5 units per port (15TB) in SCSI mode or ten units per port (30TB) in Fiber Channel mode. They also sell a massive backup storage server that holds up to 6GB per unit in 5 rack units.

Aberdeen sells their scalable NAS system that holds 3TB per 2U unit. It's expandable up to 5 units per port (15TB) in SCSI mode or ten units per port (30TB) in Fiber Channel mode. They also sell a massive backup storage server that holds up to 6GB per unit in 5 rack units.

Aberdeen has some nice rackmount servers, including the TeraStorus box that holds 24 SATA drives and a server. You can pack a whopping 6TB of storage in one simple box. Use their configurator and choose what you want. If you need custom built stuff, give them a call.

BTW, their 5 year warranty kicks ass.

Aberdeen has some nice rackmount servers, including the TeraStorus box that holds 24 SATA drives and a server. You can pack a whopping 6TB of storage in one simple box. Use their configurator and choose what you want. If you need custom built stuff, give them a call.

BTW, their 5 year warranty kicks ass.

Aberdeen has some nice rackmount servers, including the TeraStorus box that holds 24 SATA drives and a server. You can pack a whopping 6TB of storage in one simple box. Use their configurator and choose what you want. If you need custom built stuff, give them a call.

BTW, their 5 year warranty kicks ass.

mean-time-to-failure of a hard drive is 15,000 to 20,000 hours

No, much more than that. A 120 GB WD Caviar for example is rated for 500 000 hours. While these MTBF figures do not mean that a single drive would last that long, it gives an estimate of the relative failure rate of new drives. When the drive gets older, it fill fail more often, but in the beginning of its life, the failure rate is much lower than your figures would suggest. So Google hardware guys are changing the failed disks only once a day, not once every hour. And when the drives do get old, they're better off changing also working drives for new ones.

Check out Aberdeen. These are almost like build-your-own rackmount servers and are a lot better than Compaq prices.

Of course, these are only good if you are not using vendor supplied management solutions like Compaq's Insight Manager and Dell's OpenManage.

--

• Indeed, there are no cheap, commodity motherboards.
• This lack is important.

I'd love to see the OpenPPC Project do something to provide sources of PPC motherboards that don't cost thousands of dollars.

Unfortunately, what I'd really want to see is a mobo that costs only a few hundred dollars, and which allows hooking up a couple or four PPC chips. And it looks like there's not going to be any combination like that any time soon.

Remember: The Intel alternatives may not be "pretty" hardware, but they do make for a compelling lowest common denominator. I can head to Aberdeen and locate an SMP Pentium III motherboard costing a couple hundred dollars, toss in a couple CPUs, and have some reasonably powerful hardware for about $1000.

For PPC to provide a realistic alternative, it needs to either:

• Provide compelling amounts of computing power that are unavailable on Intel;
• Provide reliability unavailable on Intel;
• Provide a price that is not too much more expensive than an Intel-based system of similar power.

Feel free to s/Intel/Athlon/g or s/Intel/Alpha/g as needed.

The critical point here is that if the PPC system is outrageously more expensive than an IA-32 system of relatively comparable power, it just won't sell. There are some that are sufficiently bigoted against Intel hardware that they'll pay more for something else, whether PPC, Alpha, or SPARC. I'm not going to pay a $1500 premium to run the same code, recompiled for PPC in order to have a PPC label on the CPU that may not be visible in any meaningful way unless I put a sticker on the case.

By the way, you may not be quite right about the "only" PPC motherboard being from Motorola; Cogent Computers appears to have one that costs around $1200. Of course, the CPUs to toss into it seem to add another $1200 or more in price, so I could be off here...

I'm certainly with you in being disinterested in "buying a Mac and ripping its guts out."

I did recently an lookup about rack chassis
in Web and here are the results:

http://www.gtweb.net/rackmount.html
http://www.sliger.com/
http://www.famous-computer.com/
http://www.inco.de/
http://www.pcwmicro.com/Rackmt/compare.h tm
http://www.rackmount.com/
http://www.aberdeeninc.com/abcatg/r ackmount.htm
http://www.technoland.com/chassis.htm
http://www.ittools.com/Products/ap ollo_cases.htm
http://users.cwnet.com/fotra/CHASSIS/
http://www.vatyx.com/rackmount/rackmo unt.html
http://www.uslogic.com/igc/igcrackraid. html
http://www.dcsis.com/rack.htm

I don't know prices - go find yourself. Some of
those are pretty expensive, but there is cheap
ones also.

I prefer my local dealer's (in Finland) chassis, since those are pretty good and nearby me.
Pages are only in Finnish, sorry about that.

http://www.damicon.fi/rakki/index.html

And sorry about possible Slashdot-effect-meltdown to every site in this list ;-))))