programming and human factors
by Jeff Atwood
I was thrilled to discover that Santa Claus left a little unexpected present on my doorstep on Christmas Eve: the two Lenovo ThinkServers that I ordered for stackoverflow.com! They weren't supposed to arrive until sometime next week.
I immediately began unboxing the servers with all the eagerness of a kid unwrapping his Christmas presents. The servers are barebones, with basic levels of CPU and memory; I bought some hard drives and extra memory to have on hand for testing and installation. Configuring servers on Christmas Eve = the best geek christmas, ever!
Oooh. Just take a gander at all that hot, sweet server hardware.
After carefully unpacking everything and taking an inventory, my heart sank.
These Lenovo ThinkServers don't include any drive mounting brackets. Which means I can't install the hard drives. What's worse, there's no way to buy the drive brackets alone; you must purchase Lenovo's "server" hard drives if you want the mounting tray / bracket assembly. And Lenovo's drives start at $100 for a generic 160 GB SATA hard drive. That's a heck of a premium to pay for a drive tray. And I'd need eight of them. For comparison, I paid $80 each for a set of 500 GB SATA server class hard drives.
I had naively assumed that these servers would come with the necessary drive trays, just like they have slots for memory and CPU. Or at the very least the drive trays would be items I could purchase individually. In the case of the smaller 1U server, I can prop the bare SATA drives into position by placing a thin book under them, which is OK for test purposes, but hardly a long term solution for a server I need to ship to a data center.
It's amazing how quickly I went from best geek Christmas ever to worst geek Christmas ever. All for want of a few lousy, stinkin' hard drive trays! It's engendering some serious Nerd Rage.
I guess I'll be either returning these Lenovo ThinkServers, or selling them on Craigslist. How sad to see perfectly good hardware go to waste.
Update 1/11/09:
I bought two official $100 drive rails from Lenovo. Pity that they come with worthless 160 GB hard drives attached. Oh, and as an extra bonus "up yours" to customers, they use Torx screws.
Thanks to some eagle-eyed Coding Horror commenters (seriously, you guys rock), I also found an eBay seller with slightly older IBM SATA removable drive rails for sale at $25 each:
The older IBM drive rails work perfectly in the newer Lenovo servers, although the front design is ever so slightly cosmetically different. The model # is IBM 42R4131, and they're for the older IBM xSeries 3250, x306m, x3550, x3650, 3800, 3850 servers. So the good news is I only have to buy $250 worth of drive rails, instead of $1000 worth. Or at least that's what I'm telling myself..
L M A O
sheesh on December 26, 2008 5:29 AMThe term 'shooting oneself in the foot' comes to mind.
I try to figure out what it is that Lenovo is trying to accomplish here, but I can't decide whether they're evil or incompetent. Incompetency is assumed, but really, could even marketing people be this stupid? This is the kind of thing which even the CEO of a large company could understand and get angry about, which generally means that for an IT techy it's possible to make the company buy from another vendor next year.
I guess this is why more and more companies just go generic, or even put their own systems together :)
Maya Posch on December 26, 2008 5:34 AMThis was very predictable. If you go back to the comments on your original buy vs. rent blog post, there were several people who recommended pre-installed, expensive SAS drives. When I saw your total expenditure for these boxes I thought, there's no way that's enough money for what you need. Just bite the bullet and buy a box with 6 3.5 SAS drives, as big as you can get them at 15k RPM. Don't cobble the system together with random parts. These things need to be tested as a unit, with hardware and firmware on the RAID controllers and hard drives all agreeing with each other.
Eric Z. Beard on December 26, 2008 6:31 AMIf you think Jeff has a problem with just rigging it up to work, then you must not have seen this website of his called StackOverflow
[RIMSHOT]
...kidding, kidding...
HB on December 26, 2008 6:42 AMLOL... just... LOL
A few geeks raving and ranting on about how we will never buy our 17 servers from Lenovo again !!11!. I bet they're shaking in their boots big time ;)
Surprising how few people here actually got the Lenovo doesn't care about or even want customers who only buy 10 machines point. Most people here don't seem to get that a real enterprise datacenter works a little differently than their 10 rack geek playground.
someone on December 26, 2008 7:03 AMI dismantled an 8 drive lenovo server a couple of weeks ago. I believe the scrap parts (intended for the trash) are still at the office. I will check at the office today, maybe they would be interchangeable. They do look identical.
Nathan on December 26, 2008 7:08 AMAh, duct tape and zip ties, my friend. lol
Craig on December 26, 2008 7:11 AMYears ago I used erector-set parts to mount a hard drive. 4 holes were drilled into an emptyish spot on the case.
iNFiNiTyLoOp on December 26, 2008 7:46 AMJeff,
As many said--SuperMicro is your friend here. You can get redundant PSUs. You'll have to put in your own RAID card, but that's what you probably should have done with the Lenovo anyway.
There can be big perf differences--at least for RAID5 or 6--so this is one area you probably DON'T want to skimp and rely on what Lenovo would have supplied. (The perf differences may not be too much if going the RAID 10 route, but then a cheapie onboard RAID from SuperMicro would probably have been just fine since there's no parity calculation involved).
If you're going for a large GB array, I'd still suggest RAID 6 for simplicity and rebuild safety (db perf purists be-damned).
TorgoGuy on December 26, 2008 7:55 AMJeff -
It's slightly amusing you're stymied by the lack of a few pieces of bent metal. I doubt a dud BIOS flash would impede you in the same way. Go to the basement, select sheet metal stock of the appropriate thickness, cut to size, put into your metal bending machine, drill necessary holes, and you're done.
If you don't have a basement (etc.), there's someone in your neighborhood who cuts and bends metal either for a living or as a hobby. The cost should be: not much.
How did you cope in the days of soldering and wirewrap?
I'm going to second the SuperMicro idea that several others have talked about.
We have probably 50 or more supermicro servers and can easily recommend them. As long as you buy the right motherboard for the case (or just buy it bare bones) you should be good to go. The build quality is just as good or better than most of the other OEMs (Dell, HP, etc.).
We haven't had to use customer support many times but they were always helpful when we did. Once we had a raid card that wouldn't work with one of their motherboards and they gave us custom bios builds to fix the problem until they could release a new official bios. Try getting someone like dell or lenovo to do that. (Not that i have add a problem will dell in the past, but getting a custom bios doesn't sound like something they would/could do).
Tony Bunce on December 26, 2008 8:55 AM** Batteries not included
Mauricio Pastrana on December 26, 2008 9:38 AMAs an insider, I would not recommend Azure. From my interactions with those fellows, the group seems to be a huge mess. Way too many architects, no clear vision, no business plan, way too many senior (principal in Microsoft lingo) people who think they're smarter than everybody else on the team, and, as often happens when all these ingredients are present, version 1 is WAY, WAY too huge in scope. As a result, it will only become usable by V3 SP2 (or whatever the service counterpart to it would be called). The only way to build a large system that works is by first building a smaller system that works. Someone needs to beat this simple wisdom into their heads.
Anonymous on December 26, 2008 10:53 AMWhy not fabricate your own? If not from measurements, then possibly buy one from the manufacturer and then copy them. I see you're in Berkeley. I wish I was only 35 miles from a TechShop. http://www.techshop.ws/ But they keep postponing opening one in Austin.
mTIE on December 26, 2008 11:02 AMJeff, I may have some of these on hand that come from failed HDDs. If you're interested, let me know!
department_g33k on December 26, 2008 11:40 AMThe comments indicate that these drives have an extremely high failure rate
Not to disregard the statistical significance of three random newegg.com users, or anything, but you'll find the very same carping for every drive manufacturer and model. The only people that care enough to post anything are the tiny percent with failures on their hands. For more actual data, try here: http://www.storagereview.com/map/lm.cgi/survey_login
Third, how much of your time have you already spent on this issue? What is your time worth?
Well, I'm on holiday, and doing this essentially for fun, because I enjoy it. The usual rules don't really apply in this case.
If you want to use generic hardware, I recommend you buy your servers from somewhere like Supermicro.
Once I factored in the true hardware raid and (for the RD120) the dual redundant power supplies, the SuperMicro stuff actually wasn't that great of a deal. But I agree it's incredibly nice not to have to deal with the drive caddy used car salesman routine.
Jeff Atwood on December 26, 2008 12:15 PMThe fact that you're talking about running the site with just two servers is slightly scary.
See the articles I wrote on plentyoffish.com , which runs a #13 US site on five servers and a MSFT stack, with probably two orders of magnitude more traffic than we get.
You've got zero redundancy.
How is this any different than my current rental situation? I have zero redundancy. I'd have to re-provision a server from scratch (beyond the OS, I mean) and restore data backups myself if there was a catastrophic failure.
Jeff Atwood on December 26, 2008 12:19 PMHow is this any different than my current rental situation? I have zero redundancy. I'd have to re-provision a server from scratch (beyond the OS, I mean) and restore data backups myself if there was a catastrophic failure.
One step at a time. Next thing you know you will be purchasing a SAN!
Dustin Laine on December 26, 2008 12:26 PMYou have been nave, indeed.
Most hardware vendors want you to buy their own drive mounts...
I did find this shop in Texas: http://www.swt.com/ they do great machine. But it is difficult to get them in Belgium (shipping, question of waranty, insurance,...). does any of you got some good shop around that doesnt charge your first-born soul for the priviledge to buy from them?
DomreiRoam on December 27, 2008 2:10 AMwhy do simple coders think they have knowledge of hardware?
Wolfgang on December 27, 2008 3:37 AMGO TO SUPERMICRO.com
They have the BEST servers, hands down. They even have a 1U server that actually has 2 seperate servers, compressed into 1U of space. They really do amazing stuff.
If you they are trying to lock you into their hardware (VENDOR LOCK IN) then screw that.
Ive bought 6 supermicro servers and have NEVER had issues as u described. So, look for good supermicro servers and websites that sell them. As always google the seller to and read reviews and look to see they are not a scam site. Get yout Barebones system, mem, drives, cpu and you should be good.
I hope everyone knows that what Jeff atwood got vendor locked IN, and this is not acceptable, by any standards.
Good luck and happy holidays.
Mayy Kukowski on December 27, 2008 3:39 AMlol, so apple aren't the only scumbags with a proprietary tax?
fred on December 27, 2008 3:40 AM@tony
Oh, we talked to IBM that a new suse linux installation does not support their raid any longer (but IBM had proposed to use 10.x instead of 9.x). Nearly 4 hours later (21:00) an IBM engineer was at the machine and installed a new driver on the first of many machines...
It all comes done to service level agreements in the end. And how valuable you are as a customer. IBM is good if your time costs more than a bit of money.
Artor on December 27, 2008 3:47 AMWelcome to the wonderful world of PC servers. You got near-component-cost pricing on the server case, motherboard, powersupply and such, but IBM had to find a profit somewhere in there. They did it via the drive sled trick (they'd also have been happy to sell you memory and CPU parts at a fairly high markup as well). It's not uncommon; all the big guys do it.
That's how they finance the RD that went into the case, mobo, management software, and support infrastructure which differentiate this system from a $700 tower system of similar specs. And it does need to be different; the tower system would plug into one enthusiast's desk. The rackmount server system must plug into a datacenter which has very different operational requirements.
Sucks that you had to find this out the hard way, Jeff. As many others have pointed out, SuperMicro is the compromise solution for those of us who aren't putting up dozens or hundreds of systems and amortizing them out over 5+ year lifespans.
quux on December 27, 2008 9:30 AMNot to be mean, but this is what you get for choosing to buy and doing it yourself with no prior experience.
Anonymous on December 27, 2008 10:06 AMJeff, just fabricate your own. As you said yourself...
a href=http://www.codinghorror.com/blog/archives/001185.htmlJust whip out your trusty dremel (you do own a dremel, right?)/a
Sprogz on December 27, 2008 10:44 AMObviously I can't even follow _simple_ instructions.
No HTML, OK.
http://www.codinghorror.com/blog/archives/001185.html
Beginners mistake.
Todd on December 27, 2008 12:41 PMSame Problem as I had with HP. As I can see in the comments - all big server-firms (Dell, HP, IBM, FSC{?}) do the same: selling overpriced hdds in their trays
Schnacker on December 27, 2008 12:55 PMAnd this is why google and such used commodity hardware.
daved on December 28, 2008 4:38 AMThis is an easy fix, and you're not a true geek unless you can find the part number. But I'll spare you and post the part number on Monday when I'm back at work. We buy these chassis all the time, and then stock them up with our own disks by purchasing the separate drive caddy from Infiniti Micro or another supplier.
Easy fix on December 28, 2008 11:14 AMI'm surprised nobody has mentioned Amazon Elastic Compute Cloud (EC2). You get complete control over your VM, as much or little machine as you need, Windows Server is supported (I don't want it, but I suspect Jeff does), EBS is on a SAN, and if your machine goes down, it should take you about five minutes to get it back up. It's not even more expensive than other less capable hosting providers.
Let someone else worry about the hardware; you worry about your app.
Marc on December 29, 2008 3:00 AMI don't think the only difference between consumer and
enterprise stuff is simply the price, but additional QA testing.
There's a bit more to it than that. Some years ago I got to play with a bunch of spare IBM 0662 1GB SCSI drives (that tells you roughly how long ago it was). Most of them were generic PC drives and one was an RS/6000 server drive. All were exactly the same model number, type, date of manufacture, everything else, the only difference was that one was intended for generic PC use and the other for server use (and presumably cost 5x as much).
Running some sort of performance benchmark on them (and again I can't remember what it was, too long ago) all of the drives except one performed at about the same level. One drive had 2-3x the performance on whatever synthetic benchmark this was. No prizes for guessing which one.
Dave on December 29, 2008 5:02 AMActually, IBM will embed specific firmware strings in their hardware peripherals to ensure that third party peripherals can't be used with IBM servers. I have personally experienced this in the pSeries and iSeries servers with both disks and RAM. However, those peripherals will work with any other equipment accounting for bus and interface compatibility. I used to use 9GB SCSI disks from an AS/400 in my home PC. Fortunately, IBM doesn't use this vendor lock in behaviour with the xSeries.
Anyways, I promised the drive caddy part number so here it is:
IBM 42R4131
Google the above to buy them for $35 each.
While you're at it, I suggest the following upgrade to get some decent performance from those disks.
Since the on-board RAID controller has pathetic performance replace it with an LSI to get great throughput and battery backup:
LSI Megaraid SAS 8408E
You'll need the SAS backplane to card cable which Adaptec can supply:
Adaptec ACK-L-mSASX4-SASX4-0.5m R
SFF-8087 to SFF-8484
0.5 meter.
Make sure that you use 15K RPM SAS disks, instead of 7200RPM SATA disks. Yes, you can use SAS or SATA disks in these chassis.
Easy fix on December 29, 2008 7:04 AMJeff:
Even if you *could* get the brackets for these, IBM/Lenovo (as of I think the X41 and later) began returning error codes on the boot of their laptops when you add a non-IBM hard drive. Sucks. I'd be willing to bet they do this with servers, too.
I will also say that you probably want to NOT use your own components in a server. +1 to Two Cents here. He has a VERY valid point. This goes DOUBLE when talking about RAM. I once bought SimpleTech, then Kingston RAM for my ProLiant database server about 7 years ago. 100% guaranteed by SimpleTech (and by Kingston) to work in my model of server. Crashed hard with a RAM error randomly. When I switched brands (SimpleTech to Kingston), the same thing happened a week later. All to save $2000 on memory. $2000 is nothing when it comes to your business's critical operations. DON'T SKIMP here. This is the bitter voice of experience talking -- and the voice of a man who's built his own whitebox computers since the mid-90's.
Servers really are a different animal.
Dave Markle on December 29, 2008 9:15 AMAs someone who had worked in the enterprise storage industry for many years, I'd like to offer my take on why storage vendors try to ensure you use their hardware.
Obviously, the answer is money. But not how you think.
They spend a LOT of money on hardware qualification. Many months and even years of testing goes into each and every product. In order to keep quality high and make deadlines you must limit the variations on hardware that you test. This means picking two or three hard drive vendors and a few models from each. And we found issues. Lots. Some were issues that you may never notice but are a true danger to your data. We had vendors provide us with specific firmware for their drives that fixed issues we found and those fixes may not make it into the consumer level firmware or firmware for other vendors.
You may say So what, let me just use my drives anyway, I'll accept the risks!. OK, fair enough. Problem is that when customers run into an issue, they don't say By the way, I'm using unapproved drives. Enterprise customers expect reliability and very prompt issue resolution when they do run into issues. Technical support teams engage upper level technical support, and even engineering very quickly to try and resolve customer issues. Many valuable resources can be spent before you realize that the customer is doing something that cannot be supported. I can't tell you how many times I was pulled off my current project to go and try and replicate customer issues. This often involved gathering hundreds of thousands of dollars in hardware and many hours of engineering time setting up a similar environment as the customer's. This often involved impacts to the deadlines of products under development. The true cost of this cannot be measured.
So, in my personal opinion, if you do want the freedom to use your own mixed-and-matched components, well more power to you. But do it by building your own server and you AND the those large server vendors can both avoid the headache all together.
(Disclaimer: these opinions are my own. I no longer work in the enterprise storage industry.)
Two Cents on December 29, 2008 9:18 AMHi,
To reiterate what Marc said, why not EC2? Perhaps not in this case, perhaps you have good reasons not to go that route, but in general I've come to think that it just doesn't make economical sense NOT to use EC2/S3.
I'm no expert, but for smaller to mid sized sites I don't think there's any way anyone could haggle better data rates than Amazon can. Please let me know if I'm completely misguided here. (And I'm not talking those BS sites that offer tera-bytes for 5$/month.)
And although the instances aren't super cheap, they're comparable to hosting charges. And throw in the total costs you'd incur if you hit just a couple hardware failures a year, and I think you're coming out pretty close to even.
And life is short: I'd pay a couple bucks a month to have Amazon vet, hire, and employ competent people that know what they're doing and can actually fix things promptly. If my instance goes down I can just re-instance it and I'm back. And when the entire system goes down, which does rarely happen, everyone is affected, and they have smart people working on it, guaranteed.
I'd love to be able to validate my preference for owning hardware, just because I generally like to control EVERYTHING I can, but sadly I don't think it makes much economic sense anymore with the economies of scale that Amazon can achieve.
So, if anyone could let me in on some reasons why owning servers still makes economic sense (in the general case, I'm sure there are corner cases that require ownership), I'd love to hear them.
And good luck with your brackets. (amazon ftw!)
Ah, I see this was already somewhat addressed:
a href=http://blog.stackoverflow.com/2008/12/server-hosting-rent-vs-buy/#commentshttp://blog.stackoverflow.com/2008/12/server-hosting-rent-vs-buy/#comments/a">http://blog.stackoverflow.com/2008/12/server-hosting-rent-vs-buy/#comments/a">http://blog.stackoverflow.com/2008/12/server-hosting-rent-vs-buy/#commentshttp://blog.stackoverflow.com/2008/12/server-hosting-rent-vs-buy/#comments/a
Jake H on December 29, 2008 10:31 AMJeff thanks for the post, I am going to have the same problem in 7 days. here is the solution for both of us. IBM part number 42R4129
These guys sell it for $30
http://cgi.ebay.com/ws/eBayISAPI.dll?ViewItemssPageName=STRK:MEWNX:ITitem=290188905669
This will fit both RS110 RS120
Marry Xmas
Mazin Ramadan on December 29, 2008 11:19 AMSince the on-board RAID controller has pathetic performance
Maybe on-board RAID isn't such a good idea...
BillAtHRST on December 29, 2008 12:01 PMAh, same experience with HP servers. Solved by monkeypatching using rubber, HDDs are somehow fixed inside. My own server, my own problem however - but I just don't get long-term selling strategy, however.
Almad on December 30, 2008 2:19 AMI've never felt compelled to comment on someone's blog before.
This is why you hire developers to develop, and system administrators to administer the systems.
Seriously Jeff, it can't be that hard to make friends with an admin, can it? Of course, I don't know any developers, so maybe it is...
I started to read this blog post got to the punchline then laughed at the naivete exhibited here. But, what can you do, people have to fail in order to learn.
The reason that the big OEMs don't sell drive rails willy-nilly to any fella with a credit card is the same reason that developers release software with bugs which are easily found by users, but testing never revealed: there are simply too many configurations and combinations to account for.
As an admin that worked up from the helldesk, I'd like to say that this is about the corporate drudgeons trying to free up support staff so that they can be outsourced/eliminated, but the reality is that if the decision makers cared that much to eliminate support staff, they wouldn't need this as a reason.
The real reason is probably the little secret that drives our every move but no one admits to: People are Lazy.
Seriously though, if Lenovo was going to let you buy the rails seperately, or even install them in servers, how many changes to the ordering software would that take? The assembly line? The product documentation? The sales scripts? The support manuals?
The name of the blog causes me to assume that most of the people reading this are developers, I'm surprised that no one understands the rammifications of this one tiny thing that everyone wants changed.
Nick Wells on December 30, 2008 8:22 AMif Lenovo was going to let you buy the rails seperately, or even install them in servers, how many changes to the ordering software would that take? The assembly line? The product documentation? The sales scripts? The support manuals?
ordering s/w: add a sku/line item for extra drive caddies
assembly: one more part to toss in the box/slide in
documentation: spec what size restrictions/interface are required for the drives
sales: add some crap for when people request this, or just fill out the slots with empty caddies.
support: don't support this configuration?
Try pushing it to the point of getting a RMA number. Often the returns department has power the sales (post sale support) does not care to have.
I ran through something vaguely like this dell - ordered system with XP home meant to get XP pro. All I wanted to do was pay $100 to upgrade. Countless hours later it turns out that they dont know how to charge for that and thus cant ship it. Finally the returns person gave me XP pro for free. Still a WTF, but at least it ended well.
Steve on December 30, 2008 12:51 PMDell is a bit different, you can find spare parts for most of Dell servers on ebay and many websites. I think finding SUN, Lenovo is a bit harder.
Brennan on December 31, 2008 1:04 AMThis fall while configuring a ~$3000 server for a small business, I had the opposite situation. I was wary, so I asked. The Dell salesperson insisted I had to buy drives in order to get the rail sets. When the server arrived, there was a full set of rails for every drive bay. I'd ordered the minimum of three small drives, planning to do an upgrade to larger capacity / less expensive drives. Yes, Dell's price for drives was six or seven times what I could find elsewhere. I'd be willing to consider the argument that their server drives are extra-special, but what sort of evidence can be presented that they actually are? Dell doesn't promote any facts, apart from the higher price. They don't warn you that their RAID can only be used with certain certified drives, do they? To top it off, they didn't RAID and partition the drives the way I'd ordered, so I had to re-do it all.
jfoust on December 31, 2008 11:23 AMWe use mostly HP proliant servers at work and buy rails by the dozens on ebay. As far as I can tell the rails have been the same for a decade. Many of the rails we buy only say compaq on them, as they're pre-merger!
I would assume you could do something similar with theses servers, but as you say they could be too new. But I find it unlikely they'd regularly change their rails, think of their cost to stock all the different models.
Of course we also tried to buy extra cpu's one ebay, but the those are coded - only HP processors will work, generic ones fail to boot.
Steven on January 1, 2009 9:29 AMJeff, I can see how this is frustrating, but don't yell at the disks, they don't take well to that.
http://hardware.slashdot.org/article.pl?sid=09%2F01%2F02%2F0626201
Eugene Katz on January 2, 2009 1:37 AMI don't know of any big-name pre-built servers that don't do this.
(HP/Dell/Lenovo/whatever. And I had the same with Sun, but maybe they've changed that if other commenters are right)
Sucks, but unless you can get decent server cases, PSUs etc and are happy to build it yourself then you have to put up with it :(
(I've not found any self-build servers which are price and feature comparable, so if you know of any then I'd be interested to hear :)
Hey Jeff. I bumped into a similar issue a while back. I solved it by picking up a NORCO RPC-4020 4U Rackmount Server Case ~$300 w/o PS.
(http://www.newegg.com/Product/Product.aspx?Item=N82E16811219021)
which has 20 SATA drive bays. While the drive 'sled' is proprietary, the SATA cable is not. I ran the SATA cables from a couple servers into the drive chassis. Bonus that I was able to use a separate power supply for the drives than the servers.
Case is good enough. Best deal for the money, for the size of drive arrays I was working with. Another option, if you don't get the parts.
A few more pictures of the internals.
http://www.norcotek.com/RPC4020.php
Chris on January 4, 2009 9:07 AMRemember the good old days when RAID stood for Redundant Array of Inexpensive Disks? *sigh*
The design of the system as a whole and contingency plans matter a lot more than whether or not you use approved drives. I wouldn't trust a single $80 SATA drive with my data, nor would I trust a single $800 SAS drive.
If consumer SATA drives are keeping you awake at night, you're doing something wrong.
Nick on January 4, 2009 12:54 PMNever underestimate the power of Duck tape :)
Jonny on January 5, 2009 6:38 AMYeah... I gave up on HP/Compaq/IBM years ago because of the bass ackward way they make components - usually ONLY works for their hardware. It's sort of like the GREAT car manufacturers...new model every year, with all new parts..just dandy! Anyway, not sure how complicated the drive tray is, but if you know anyone that has a metal shop, have them make them and sell them. Then, maybe the IDIOTS might get a clue for their next design.
doublehelix on January 5, 2009 10:25 AMTorx is an 'up yours to customers'? Any serious geek has a set of security Torx drivers[1] in their toolbox. You really have lead a sheltered life driving a keyboard!
1 http://en.wikipedia.org/wiki/Torx#Variants
Colonel Sponsz on January 21, 2009 9:11 AMThey're security Torx, so.. yeah.
Jeff Atwood on February 5, 2009 3:26 AMWish I would have saw this blog before I bought my RD120! Ebay it is.
Alvin on August 3, 2009 6:29 AMOr, even better - due to the massive publicity of this blog, Lenovo might be kind enough to offer mounting brackets at the price of their value as a good faith Christmas kinda deal.
Josh Smeaton on February 6, 2010 11:12 PMI guess your best opportunity is getting the trays from eBay.
The server on left looks identical to IBM x3650, so the trays should be also identical.
Look for 42R4129, 42R4131, H87282 as part numbers.
I once stumbled over the same issue myself and i had to resort to eBay.
Don Zoomik on February 6, 2010 11:12 PMThere's really two choices for a server... build your own up in a supermicro case, or go with a brand. If you go with a brand, you have to order accessories from that brand.
There are pros and cons to each solution and I would not say one is always better than the other.
Steve Sheldon on February 6, 2010 11:12 PMExactly the same situation when I bought a Dell PowerEdge T300. Out of four drive bays, it only came with two drive brackets for the existing two disks. The other two just have placeholder covers.
When I called Dell to check, I have to order disks in order to get the brackets/carriers. There is seldom an avenue to get by these cashing schemes unless you personally know somebody who works there.
Aaron Seet on February 6, 2010 11:12 PMThe link to the forums page does provide a bit more detail on the price difference: the extra money you pay is for the warranty, which provides for next-business-day replacement. The server drive may well be precisely the same hardware as a non-server drive, but the extra money is for transfer of risk from you to Lenovo.
As to why they won't sell the sleds separately, well yea that looks like profiteering.
Denton Gentry on February 6, 2010 11:12 PMSimple enough.
Here in Germany the company is Thomas-Krenn; sure enough similar companies are in other countries, too.
What they do - and what I buy:
They offer Supermicro based servers *and* (optional) 3-year service, ranging from next-day up to having their technician at your site same-day within 4 hours.
BTW: They *do* have 1 HU servers with redundant PSU.
I like that because it offers me choice, from allowing to buy naked Supermicro (or whatever) parts at another company up to a solution packet quite like HP or IBM ones for companies who need the insurance part.
Even better, being quite standard, their parts can be changed quite often. Example: Once, I wanted a special ventilator. No problem, I just replaced it, as both were 80 mm standard ones.
Another view:
There are different clienteles. Based on some 20 yrs of experience my rule of thumb is: The bigger the company/customer the more paper/legal/insurance oriented (rather than purely tech) they act.
To be fair, in those prices, e.g. for drive rails, companies like HP or IBM have to include a lot of know-how and, more costly, a structure and a pool of techs, warehouses, reps, aso. Simpe reason: To their large customers they are basically a hardware provider and an insurance company.
Good luck with your drive trays !
The comments to this entry are closed.
Subscribe in a reader
Subscribe via email
|
|
Traffic Stats |
