Slashdot Mirror
Unreliable Linux Dumped from Crest Electronics
nri writes "The Age writes, Linux misses Windows of opportunity. Crest Electronics chose a Linux operating system, then seven months on, the company chose to abandon it for Windows.
Mr Horton says. ".. the machine would basically, putting it in Windows terms, core dump or blue screen at random. It would run for weeks or so and then just bang, it would stop....I fully support Linux but if I had to make the decision again I'd pick Windows. A big reason is the fact Windows was up and running in two hours at all the right patch levels. The installation of SAP took two days on Windows, the installation on Linux Red Hat took two weeks. The total cost of ownership is actually lower in this case than with Linux because of the hidden costs of the support.""
...we will see what you have to say about hidden costs and core dumps.
HA! HA! HA!
Anyone that says that Linux will beat out Windows in every situation is a fool.
Choose the product that best suits your needs. If Linux doesn't cut it, get Windows. If Windows doesn't cut it, get Linux.
It costs money to hire qualified admins, Windows or Linux.
if you don't know what your doing a linux box can be as unstable as jello on a trempoline. I agree however that a windows box is quick to setup compared to a linux box. But the différence is that once you setup a stable linux box. You dont have to touch it ever. Except to patch, and admin. But if you documented your config correctly the minute you have to build another one. it's a snap.
What, Joe's Special Distro? All the SuSE and RedHat machines I've ever run were rock solid. They weren't running SAP, mind you, but still.....
7 November 2006: The day Americans realized corruption and incompetence weren't addressing 11 September 2001
First of all... blue screens?
OK, getting that out of the way, obviously this crew of nitwits couldn't tie their own shoes with encouragement and instructions. If I were the computing engineers involved, I wouldn't be writing any articles or letting anyone know that I failed totally to understand modern computing. Doesn't look good on the resume.
...Steve
They got to do what they got to do but my results have been much different.
My server machines regularly run for a year or two without rebooting. About that time I invariably decide it needs more memory or some other hardware upgrade that requires a reboot.
Coding Blog
Obviously, your admins were not qualified to administer a Linux server like this. If it took them two weeks to get software installed and running like that, I'd fire them right away. Even if it is SAP, a complex piece of software. Just because you got it up and running in 2 days on Windows doesn't mean it was done right, or done securely.
"the machine would basically, putting it in Windows terms, core dump or blue screen at random"
whereas you can expect windows to core dump periodically and predictably.
Installing SAP on Linux in the first place? My guess is that the second time around (on any system) is much easier than the first. How long have they run SAP on windows yet, and have they had the time to see if it would blue-screen? How much of a dependence did they place on potentially lower-level or unsupported drivers on higher end hardware where the vendor focused on Windows instead of Linux? What vendor's hardware did they use in the first place?
All these questions point out to: Not enough information to make an informed opinion on what acutally happened. Disclaimer: I haven't read the article yet, just the summary.
FWIW, us geeks know how to tweak things and get everything hunky-dory. For most users though, Linux needs to
just work". Yeah, it depends on what you want to do with your computer-- every distro I know of "just works" right out of the box, as long as the box is all you're running. No flames here, just an honest criticism. Windows does indeed just work. I see where the guy is coming from.
(For the few minutes til it gets compromised).
how? and who funds Crest? That statement they made was almost word for word from the Microsoft playbook, seems suspiscious.
Odd that the Windows terminology for the blue screen of death now seems to be the standard term for a computer crashing. Or maybe that's not so odd.
(please don't mod this as funny, I am very serious here.)
They got to do what they got to do but my results have been much different.
My server machines regularly run for a year or two without rebooting. About that time I invariably decide it needs more memory or some other hardware upgrade that requires a reboot.
Coding Blog
I wish he would have given us more information regarding the problems he ran into. I'm talking about system specs, the name and version of the Linux distro used, and more information regarding the software they apparently had so much trouble installing.
When problems do happen, the open source community is notorious for getting them fixed very quickly. If he were to provide us, the community, with more details about the problems he encountered, I just know they could be solved for him and potentially for many other users in a similar boat.
Cyric Zndovzny at your service.
Sounds like they did something wrong...
The gates in my computer are AND, OR and NOT; they are not Bill.
limit coredumpsize 0 Thats how you keep the toilet from clogging
Maybe windows IS the best choice if your IQ would make a good temperature for beer.
Dog is my co-pilot.
What is SAP? A Google search yields a company that sells business products, but there doesn't seem to be anything related to a point-of-sale system or workstation software. Is it an electronics design software?
Colin Dean Go a year without DRM
Well, if this Crest Electronics is their website they have more problems than just Linux. From their homepage:
Currently some people are having problems accessing portions of our website. We apologize for any inconvenience this may cause.
"Good things don't end with eum, they end with mania or teria." - H. Simpson
After reading the article (yes I actually read it) I decided that this article doesn't really say much about anything so I decided to not post anything.
Oh wait.... DOH!!!
It took them seven months to realize that Linux wasn't working? That's sort of wierd don't you think? Anyways I guess it was just bad luck, eh? I mean Linux doesn't really do that for me and I am not an SAP Certified whatever-majigger...
Is that sometimes bugs in say a 2.0.5 version of software that's standardized during a Linux distro's "point oh" release just start seriously affecting your company's performance. Then you try to upgrade but find a dependency on another new package as well. Some of us cannot install whatever we want on the machines. I have to spend a week or two research and prepping and thinking of every possible reason and counter-complaint about why we need the updates that I get exhausted pretty easily.
Linux is by far the best UNIX-like system out there and I love it for that. But depending on other developers who work for free and fix all critical bugs in a timely matter is, for me, like putting a $30 million dollar project into the hands of what could be a 20 year old kid, albeit a damn good coder of a kid, but someone else nonetheless.
With Microsoft at least there's a bit of accountability or someone my company can blame when our project fails.
If you "get" pointers add me as a friend (116)!
This whole article is useless without really saying what the crash was. You could have the most rock solid stable server in the world, and it won't mean much if the applications you're hosting are buggy and badly implemented. It would be nice to know to EXACTLY what crashes he was getting and why. Not just "Uhh, there were core dumps and blue screens, but with a linux blue instead of microsoft blue." I think this would be a great opportunity for an Ask Slashdot poll. Maybe he'd even post some of the core dumps.
what a load of crap.in all my years admining linux systems i have never seen ANYTHING even remotely close to a windows blue screen style crash. a user land process cannot blow away the system like that under linux. the only way this would happen is 1. bad hardware 2. idiots playing with kernel settings they shouldn't be.
either way none of this reflects on linux's stablity at all, just on the skill of the admin running it.
want another hint this is a case of a total retard running the system? "2 weeks to install sap on redhat"
even the most stable system will go bad with an idiot in charge.
If you mod me down, I will become more powerful than you can imagine....
I can understand the long install time. This is proof on one of the major flaws with linux. Poor documentation, poor standards across distros, and obscure undocumented dependencies. Don't get me wrong; I have been using linux since 1999 and have come to appreciate a lot of it. But still to date I want to bang my head on my keyboard when I install some new software and I am told that I need such-and-such lib or a different version of something. Then a good part of a day is shot trying to track all of this stuff down and get it installed. What I have just said should be tempered by the fact that I do not believe that windows is a good choice. 99 times out of 100 you have an unstable machine that costs you huge $ in downtime. This is where linux (and Mac) is good. Once you get it set up it is rock solid. I guess that you have a choice, long set up with linux then less maintenance or short set up with Windows, and a lot of further maintenance.
My
Well, from this point on I'll be skeptical of the skills of any IT person that has "Crest Electronics" on their resume from now on.
Like many of you, I will generally support Linux over Windows in every (server) situation.
However, one client of mine (Oracle E-Business Suite on RedHat 3.0) had major issues. Turned out that their problems were related to kernel version and the particular version of EMC FC/AL drivers that were loaded. This required significant troubleshooting, several kernel patches (a few "experimental") before finally getting worked out. The problem caused the servers to completely lock up under load.
Admittedly, we had a SAN architecture that is probably somewhat advanced for a Linux installation. I also acknowledge that this sort of issue could arise on any platform. However, the response we got from RedHat was less than I've experienced from other *nix vendors. (That said, what kind of response would I have gotten from Micro$oft?)
I guess my point is, given the problems we encountered, I'd consider a different platform too. (Although I'd tend towards Solaris/AIX or even, maybe, HP/UX)
he has a point. Sure linux is free, which is nice, and it is a very good selection for many companies, but there are some problems. Linux support in terms of calling can be shaky, and support is something companies need. Another problem is that it is often easier to find needed software for windows that employees are familiar with than it is for linux (Not neccessarily better software). Also, linux is far more advanced than windows. Although this does allow for a number of cool things like several graphical interfaces, it also makes it harder to figure out problems and harder to find competant people to manage your linux boxes. Lastly, although we all known how ridiculously riddled with holes and bugs unpatched Windows are, patched versions aren't always that bad. Sure, you get the security holes, but you don't get too many blue screens of death anymore. As long as you run a firewall, dont' use IE, and run a few other things like Processguard your windows is fine. However, one argument i did not like was the hidden cost. Yes, it costs more right now because you have the same IT department that was workign with windows making a dramatic move to Red Hat linux. And for a time, you do risk security concerns b/c your admins are incompetant for awhile. However, now you don't really need to pay for windows licenses, don't need to pay for firewalls (look at the end user agreements for most free windows firewalls - home and personal use only), don't need to pay for an AV, and you don't need to update Redhat as often as windows nor is it as costly. And in the long run you are less likely to get huge security breeches with almost any linux distro over windows.
what in the world could he have been doing for it to take two weeks?
And before anyone starts talking about storage arrays, databases, etc etc...those things are no different in windows. Try again.
And by that, I mean people who have experience and know what they're doing. In some ways I might argue that it's actually more difficult to properly admin a Windows box well.
I'm switching to Colgate.
That the decision to go Linux was made by his predecessor.
Looks like 'new manager' syndrome to me...
Quidquid Latine dictum sit, altum videtur (anything said in Latin sounds important)
Maybe the reason it took two weeks is because they had windows admins set up linux servers. Maybe that's _also_ why it didn't work well. Maybe they need to do some training or hire someone with the right skillsets?
Come on - this has to be a troll! Perfect title, good weighting in the blurb, nice "put it in Windows terms" and then the very reasonable "oh. but I fully support Linux"
I'll bet dollars to dimes that they didn't have a good Unix Sysadmin on board, just some guy who was trying Linux out. And I bet we see more of these "failures" when people try Linux expecting it to be a silver bullet & cry when they get burned because they don't have the qualified staff.
Damnit - I wanted my nick to be "WouldIPutMYRealNameOnSlashdot"
You can always run Linux on crappy hardware and blame the OS. I bet this way you can put two systems side by side and have the Linux one crash 10 times more often than the Windows one. Or is this only the latest twist to the "Get the fakes" campaign so nicely concoted by Mr. Taylor of M$ ?
I am Joe's enraged, inflamed Linux installation.
When a program dumps core, it means that the program did something that it wasn't supposed to do (like try to read memory that isn't valid) and the operating system has (correctly), stop the program's execution, and to make life easier on developers, copied the program state into a handy file so that the problem can be debugged. No other programs on your system will be harmed by this one malfunctioning program.
When Windows blue screens, it means *the operating system* has done something it wasn't supposed to do (like try to read memory that isn't valid) and the operating system bails. Often, it will return execution to the next instruction and hope things will be okay. It almost certainly isn't. You're basically screwed.
The equivalent in Linux is an Oops. They don't happen that often on production systems. A crappy properitary program doing things it's not supposed to is *not* a Linux problem nor an Open Source problem. It's SAP's problem.
This is a testimonal about the crappiness of SAP and nothing more. They obviously didn't do enough testing on Linux.
It took them two hours to install and patch Windows?
SAP is a Big-Azz database/inventory_control/accounts_payable/accoun ts_receivable/purchaseing/accounting/blah_blah_bla h monstrosity that makes my life HELL!!!! I hate it I hate it I hate it!
It is akin to Powerpoint, makes more work but looks glitzy! (at least our implementation is)
Your thin skin doesn't make me a troll
"The Best Run Businesses Run SAP" is a true statement... SAP says it over and over again. What they're really stating is that only the best run businesses can survive a SAP implementation, the rest run out of money or patience, or worse, end up being driven out of business by the enormous cost and disruption it causes. SAP has a HORRIBLE track record on linux. They claim support for linux and other non-MS platforms, but that's only for their core products. Everything outside of CRM and R3 is riddled with technotes and disclaimers about needing MSSQL and WINDOWS. They don't really write cross-platform systems, they just make claims and back them up with fine-print disclaimers.
I just left a company that was $10M and 2 years behind on their "$2M" SAP implementation. It's a joke. Once SAP gets their foot in the door, they flood your company with incompetent consultants and rebuild your business around SAP-approved procedures and architecture. At the end of this clusterfuck you end up WAY over budget and desperately looking for a scape goat. Clearly Crest Electronics chose Linux.
SAP products require patch after patch, and take MONTHS to really install. We had a team of engineers working around the clock (literally) for 5 months to get our base systems set up to SAP specs. Even then we would receive "mystery" patches, frequently resulting in system crashes as they weren't designed to work with other patches. Bottom line - SAP is the problem. They churn out highly unstable software and have armies of consultants who will sweep problems like this under the carpet or find something else to blame.
The article states clearly that it was Red Hat. Over and over and over and over...
Blue screen is a Windows thing but core dump is not.
Crest Electronics is trialling Microsoft's Windows Server Update Service, which allows automatic patching for the operating system and other Microsoft software on servers and desktop machines across a corporate network. Its benefits are one of the key reasons why Mr Horton stands by his decision to switch from Linux to Windows.
"We run Linux on our web server and for an accounting package with great success and we do use the auto-patching in those environments,"
I work in a Windows shop but we don't do automatic patching. We don't patch until we've done extensive testing on our own to make sure it works in our environment first. SUS/WUSS/whatever is great in the sense that it allows you to control how patches to your Windows workstations are distributed. You can change the workstations' auto-update behavior so they only update from your SUS servers, etc. But the automatic update thing, from what I've heard, is rarely used in a production environment. In fact, Microsoft gives you a considerable amount of control over its behavior, probably because in recognition of the dangers of auto updating in a production environment.
Mr Horton disagrees: "It might be fine for things like security patches, which don't impact SAP certification rules but with some patches you still actually have to check the release levels and then check against the SAP site. Otherwise SAP might ask you to roll back to the previous version before they will support it."
Give me a break! The same thing happens in the Windows environment. It took Bloomberg and our other vendors a while before they supported Windows XP SP2. When SP2 first came out, a lot of vendors blamed SP2 for problems that may or may not have been SP2's fault. It took Windows vendors a while to adpot SP2 as well.
In any case, the whole patching issue he takes with Linux seems absurd. Just a few days ago, I think our server guys patched their cluster with a Microsoft service pack. Now the cluster refuses to fail over properly. Patching in a production environment is ALWAYS a big headache if you want to do it right. Unfortunately for our server guys, we don't have a spare cluster sitting around for them to test patches on like they normally do with other servers.
EvilCON - Made Famous by
They said that the Linux box was running for weeks without a crash, but took two weeks to install SAP. The Windows box would be configured in about two hours, but how long would it run without fail. I've never had a Windows box that would run for more than a month without fail.
Does the purchase price of Windows, plus the greater downtime (cumulative) make it inferior? All that has to be done is to write a script to reboot the machine every month or so and plan for this. Windows now reboots more often than the Linux box, resulting in more downtime.
Silence is golden... and duct tape is silver.
Perhaps the whole root of the problem is that, fearing to let their (future) IP out of the bag, they prevented their techs from going to the community with questions.
Yes, because we totally believe that you came up with that arguement on your own. "Total cost of ownership" is a natural concept which simply develops in natural language, like swear words based around bodily functions.
It's been a long time.
From TFA , a quote from RedHat support regarding Crest...
"We asked the customer to do a diagnostic test and the customer never responded, so it was impossible for us to address the issue," Mr McLaren says.
These Crest guy's didn't even have the ability to use support properly.
and
"We run Linux on our web server"
The entire company has 1 webserver? Unless he was missquoted this guy doesn't have a clue what his IT department should be doing.
Nuff said.
... how they think Windows is the solution when they were unable to diagnose the problem?
Our company just installed the same image on two identical certified Linux-compatable servers for the same job. One will core dump, seemingly randomly, but often enough to render it unsuitable for production, while its counterpart other runs flawlessly 24/7. Would I be way off the mark in thinking that our problem is a subtle hardware defect in one machine, not a deffective OS, particularly after trying several different kernels/distros/versions/etc etc?
While the article is light on the scale of the operation (I've never heard of Crest Electronics), it frequently uses the pronouns it or the machine (as opposed to the plural) in reference to the server(s). It would seem silly to me to consider a platform shift and the associated costs when you're having trouble with a single unit.
Quite a few respectable companies will put your precious project in the hand of the cheapest possible labor they can find, very often workers that are incompetent, lack training and have no real interest in the product they are making.
I don't give much for the argument that corporations would do a better job developing good quality software, instead in my experience cutting corners to save money seems to be way more important to the bean counters that control most corporations.
-- This SIG was handmade in a sweatshop
GOOD! If they are dumb enough to run SAP, they are the perfect candidates for Windoz.
Your thin skin doesn't make me a troll
is how Windows BSOD's all over the place because it turns out their hardware was crap.
How many passes did memtest86 successfully make? How many times did you successfully do a total system stress test? (Like rebuilding a kernel or XFree86.) I've got $5 says it was defective hardware and what they now have is SAP/Windows silently corrupting their data.
What did it do when you installed it under VMWare? You did try to run things under an always-the-same virtualizer to control the variables, didn't you? Hello? Is this thing on?
The guy saw the BSOD X Screensaver and set it to go off every week.
I can picture it now.
Blue Screen?
Kernel Panic?
Mac Crash?
Amiga Crash?
Comodore64 Crash?
WTF!!! Linux is the most unstable OS out there...
The install disk was upside down. Since I'm in the northern hemisphere I had to turn my monitor upside down to read the story from the Australian news site. The Linux distro came from the northern hemisphere which is right side up. So turn it over and try again.
What's the big deal? Use the right tool for the job. Rob
The person giving the references in this article did not seem to be the long time UNIX user he claims to be.
first: He put his experience with Linux into a windows context, suggesting that he is in fact an experienced windows administrator.
second: he did not understand automatic updates. A feature which is and has been available on many linux distro's for quite some time, and a feature which is quite prevalent in UNIX especially from IBM
third: Red Hat Linux (even enterprise class) does not have a very restrictive hardware requirement, and the odds are pretty good that they would have needed to do the same hardware upgrades to run whatever windows system they eventually moved to.
fourth: Anyone who is an experienced administrator knows that the core operating systems are tremendously stable, be it windows or Linux, or UNIX, and that the instabilities in any system will be introduced by drivers needed for operation of application specific hardware (for example a custom cash register based peripheral or some such). This tells me that they had just such a piece of equipment in their systems, and that the vendor of this hardware did not supply working drivers. Further, I would conjecture that said supplier probably had a long standing windows driver, and had ported the drivers to the linux platform specifically at the request of this client. The result is what you would expect: a first generation driver which fails intermittently.
-=Geoskd
www.geoskd.com
I wish I had a good sig, but all the good ones are copyrighted
A lot of /.'ers, when they hear of similar stories about Windows just nod their heads sagely and say "See, I told you so. Windows is a heap of crap!" This just happens to be one example of what, inevitably, will be many failures of Linux as it becomes more popular.
I am sure that if a sane and non-polemic review of the situation was performed the major fault would lie with "management". In my many years in the IT industry I have seen a few disasters and they are all a result of misguided management.
Two roads diverged in a wood, and I - I took the one less travelled by. (Robert Frost, 1916)
These kinds of problems happen from time to time: some particular combination of hardware and software will just be flaky, and if you change the OS, it all works. In my experience, it's more common to have problems with Windows that disappear when switching to Linux, but I have also (occasionally) seen the opposite.
Anecdotes like that tell you nothing about your expected TCO. Furthermore, while I sympathize with Mr. Morton's short term unhappiness, nothing supports his conclusion that there is some intrinsic flaw with Linux or that Windows TCO is lower. In fact, whatever reasons he had to choose Linux in the first place have not been invalidated by his experiences. If he thought Linux was the best choice before his string of bad luck, it still is.
Does anyone else find it odd that there's no byline for this story? Other articles on The Age actually list an author, but not this story.
From the article:
:)
"You have to be using the right certified components, otherwise SAP
won't give you the support. To go through and match everything off was quite
tedious," Mr Horton says. "After doing all that, we
came to a very interesting situation where the machine would basically, putting
it in Windows terms, core dump or blue screen at random. It would run for weeks
or so and then just bang, it would
stop."
I want to know what "components" he was referring to, (Don't worry, I know how
to use google), is it outside the realm of possibility that a mismaching of
libs, apps, whatever, even having multiple libs can cause apps that depend on
them to crash, I have no experience with RedHat, but I've seen quite a few
people use "rpm -f" (or whatever the --force switch is) instead of admining a
system properly. I've also seen quite a few windows admins treat a RedHat box
just like windows, which, if you're an admin, you just can not do. I'm sorry,
it's two different worlds.
I have never seen a linux box, after being configured correctly, just stop.
I've ripped out SCSI backplanes from machines and replaced the innards with
SATA drives, the box didn't even support booting off of a card, which was
gotten around by using a floppy, and that box has stayed up for over a year,
(once it was on a UPS) with heavy load. Something got borked on that box,
plain and simple. Maybe he should have used Slackware
SealBeater
-- Its survival of the fittest...and we got the fucking guns!!!
This sounds an awful lot like a hardware problem rather than an OS problem. It could be bad RAM. That said, a Mac has low TCO and good reliability...
From the FA:
In other words, "Crest decided that our software was at fault before consulting anybody's support staff, and refused to consider any other possibility."
I only mention this because we had a client that insisted that our server-side components were at fault for their transaction server failures. One calendar year of complaints and debugging later, the server's disk controller failed, they switched over to an identically-configured redundant server, and our software magically became as stable as the Rock of Gibraltar. Only in the post-mortem did they discover a bad memory module, and since the server was dedicated to our components, it was either us, or the OS vendor. Since we are significantly smaller than the OS vendor...
It sounds like a hardware or driver problem. Too bad their vendors and support providers failed them. Such important servers need to work 24/7. Cost and OS are pretty much insignificant in this case. If by some fluke switching to Windows cures their instability, they just need to run with it, or incur more downtime costs. I've never encountered these hidden support costs they mention, or Linux instabilities on a server, and Windows fails me all the time (never unstable, but always unreliable), but clearly it's a different story with their hardware, applications, and support staff.
All this recent talk of cost of operation differences between linux and windows is complete nonsence. Sure, to an en-experienced linux admin, properly configuring linux and setting up everything to work smoothly may be quite time consuming and therefore seem to be quite expensive to the owner of a company. On the other hand, if an experienced linux admin is running the show, everything should run smoothing, efficiently and be rather cost effective. The linux distro also has a bit of influence on the cost. Thru personal experience I've learned certain distros are harder and more time consuming to setup, run and maintain. Whereas other distros, or homebrew distros can be extremely easy, and again, cost effective.
I thought it was on OpenBSD, or some other dead software.
He talks about installing SAP like it's a file server or something. SAP is a monster (there probably isn't anything bigger in the software world) and doesn't run on just one box in ANY company. This sounds like a fake. Blue screen my ass.
They also comment that the Redhat update system is "unacceptible" because SAP may not support a particular patch, then comment that Microsoft's similar system (which does the same thing) is a key reason to change. HUH? If SAP hasn't stated they will support a MS update vs. a RH update, who is at fault when running a "certified" operating system with a standard means of performing updates as part of the OS.
Exactly. I smelled a rat the moment I read that (according to the article) SAP supports Windows autoupdating, but not RH autoupdate.
Personally, this feels like a "SAP supports Windows as a server OS better than they support Linux" statement vs. a Windows vs. Linux argument. In situations where you are buying servers just to support a single app, the golden rule is "call the support line, ask for a level 2 engineer, and ask them if THEY had to support an install, what OS would they use". That will give you the answer that in the long run will save you money.
For that matter, I'd be curious if the customer had even asked the SAP sales rep what RH and/or Windows server features they supported, or if they could supply any customer references. Sounds like someone (not the interviewee) made too many assumptions before cutting a P.O.
Luke, help me take this mask off
The total cost of ownership is actually lower in this case than with Linux because of the hidden costs of the support.
I think he meant to say: "The total cost of ownership is actually lower in this case than with Linux because of the hidden costs of the support when supported by idiots like us."
I don't care what platform you choose, if you don't know what the hell you're doing, it's gonna be a pain in the ass as well as the pocketbook.
This sig rocks the casbah.
How come it took me 2 weeks to install Windows 2003 SP1 terminal server with Office 2000? I went through 7 Microsoft Techs before I finally got the problem resolved. Granted normally it's not usually that bad, and it only takes 8 hours to PROPERLY set up a windows server...key word is PROPERLY. And how come it takes me 3-6 hours to setup a Fedora server the way it needs to be for file sharing, web, mail, whatever including updates. Seems to me that the reason they had such problems is they didn't have someone on staff that understood linux. It's not something that someone can just jump into. For that matter (using the same phrase again...maybe there's a point here as well) to set up a windows server PROPERLY it takes knowledge as well. More often than not someone that just sets up a windows server with the defaults is leaving the box(es) open to attack(s). It's all about the proper tool for the proper job. And in this case either OS could be the proper tool, but also having staff that understands the stuff on hand is a proper tool as well.
I do believe that Free/Open SW is largely better than Closed.
However, what I meant was that the accountability factor and sometimes the openness can harm you if you depend intensely on certain Open SW packages that may have critical bugs. For example, my customer will not purchase my $30 million software package if it contains critical bugs. They routinely browse bug reports and security bulletins for bugs in the packages we use. If they find one we must fix it before shipping.
I say this from experience -- Apache is hard to learn in 1 night to fix a security bug!
If you "get" pointers add me as a friend (116)!
This is sooooooo silly. This is one of those cases where they have idiots implementing the system and / or troubleshooting a problem.
I've implemented countless systems thru the years from VMS, HP/UX, AIX, Solaris, Linux and Windows. I've also implemnted SAP on Windows and Linux.
I've had PLENTY of problems with Windows and literally no problems with Linux. I never got middle of the night calls on Linux, yet was always apprehensive about those calls for the Windows boxes. It takes VERY FEW people to run / maintain the Linux boxes and MANY people to maintain the countless windows boxes it takes to run similar systems.
If a system were running a heart-lung machine, would you want it to be running windows?
This guy is a fool or a liar or both.
I recently spent a good long week or three while trying to set up a new dual Opteron server which couldn't recognize the RAID card. This isn't even anything to do with GUIs, mind you - this is simply getting Linux to see the hard drives. The big problem, in a nutshell, is hardware drivers. Hardware support remains one of the big problems for Linux, and I don't care how much you love Linux, I guarantee you'll be swearing at it before too long if you try installing on any kind of non-run-of-the-mill box. Windows, like it or not, generally has drivers. You even get them on nice convenient CDs with whatever cards are causing the problem, which is doubly frustrating. I tell you, the headaches I've had setting up Linux are no joke. It seems like every time I have to do it (every couple of years or so currently), there are a whole bunch of new little gotchas to waste my time for days. And the knowledge (once you finally figure out the magic keywords to google for) isn't even useful or transferrable - it's probably something you'll never encounter again, ever, in this form. The answer is generally some obscure incantation in some god forsaken configuration file three levels down in some /etc directory you didn't even know existed. I've had kernels refuse to boot because devfs was enabled. I've had kernels refuse to see a storage device because some other totally unrelated module happened to be enabled. I've had to go and patch the kernel code with some obscure fix to make the thing work under 64-bit mode. I've battled endlessly with weird errors that give no clue as to their source.
Ok, ok, I don't hate Linux. It's actually rock-solid most of the time, once you get it dialed in. But I just have to say that I have the utmost sympathy for anyone who jacks it all in and goes with Windows - particularly for a desktop environment. I mean, come on - Windows 2000 (never used XP) is pretty rock solid, hardly ever crashes on me and you can get drivers for all the hardware without any hassles. Anyone who says that Linux beats Windows in this regard is just delusional, sorry. And on the desktop, Windows "just works" for most users. And it has all the software they expect, and it works the way they expect. And as for all the viruses and malware - that's an education problem. You could fix most of the Windows issues out there by using Firefox or Opera, and any reasonable email client but Outlook Express, from behind a Linksys cable/dsl router or any other firewall.
Ok, sorry, had to get that off my chest... I just get so tired of reading all these pious comments putting down anybody who chooses Windows over Linux. Get real! Linux rocks, much of the time, but come on - when it sucks, it sucks big time.
I hate those *false* links that redirect to a registration page. Even if it's free, do they imagine anyone is going to fill those long forms for every page they visit.
Fortunately, the bugmenot bookmarklet did the trick.
About the story : so we have *one* situation where a problem happenned between SAP and linux. That kind of conflicts happens all the time in IT. Either you solve it or you change one component.
In both cases, drawing general conclusions on the abandonned product is common but unfair and a sign of lower qualifiquations.
Men are born ignorant, not stupid; they are made stupid by education. Bertrand Russel
They probably had staff only trained in Windows so when it came time to keep it up they had no idea what they were doing. I work with people who know People Soft through and through... on Windows... so when it comes to Unix, they are rather frustrating to deal with... I would rather be a postmaster for an open relay than deal with these people.
Argh, now I see it. Part of the text below the picture...that's different from the other articles. Nevermind.
2 hours sounds like way too long to me. I'm a tech at a linux cluster company. We can install basically any linux distribution (which we've typically included the latest updates in already) on basically any hardware in under 15 minutes.
While it is true that MCSE's come cheaper than a good unix/linux admin, you typically need more of them to do the same job. In my job, I administer all the servers, keeping on top of security updates easily. This is all in addition to my regular work(95%+ my time) installing and configuring clusters. Doing this same work with Windows would require more people than we have, actually costing more total.
Two weeks to install a Linux distro ... even compiling every thing like Gentoo would be quicker. A real Linux Admin or power user could have a distro up and running in about 30 minutes or less. Then maybe 5-30 for patching depending on your bandwidth.
I like how people blame the OS because they do not know what they are doing ... any OS will be unreliable when people who have NO clue what they are doing are administing it.
While I know that the people from Crest have no clue what they are doing.
Michael
Linux: For those able to think out side of a window
I hadn't installed linux in 2 years. bought a copy of mandrake (now some othername) had it installed and full "LAMP" configuration within 3 hours, including many custom network configurations. Its been a stable server for me ever since. We installed on a machine at work within a day (and these are hpux/solaris admins who hadn't used linux before)
Generally Unix/linux is slightly harder to admin than Windows, but not that significantly. Your admins need to be pretty good though and thats the key. The one linux failing in making admin easy is the different tools each vendor ships (YAST for Suse, Red hat has another tool, Mandrake yet another). If it was more common across linuxes administrating would be easer. Difficulty is one of the prices of choice, as there is less documentation for each choice.
I know SAP implimentations take years of planning from experience (I had a database of my creation compared with SAP, SAP had everything and the kitchen sink, but mine was faster to install, easier to use and cheaper by about over 1000x and was used till SAP was finaly implimented 2 years later)
I'm suspicious of this article because of all of the buzz words. It's practically rife with Microsoft's comparison campaign phrases. Furthermore, that SAP had difficulty running on Linux is a SAP issue, not a Linux issue.
So much for that new self-brushing toothbrush. Lets see Windows XP Home Automation do that!
The it person came in for the change and had a background of SAP on AIX.
RTFA or shut up.
Men are born ignorant, not stupid; they are made stupid by education. Bertrand Russel
I've had SAP up and running on Linux in under six hours, starting with a blank box, no OS installed. Oracle takes a bit more than eight hours. Using Enterprise Linux, it takes perhaps another twenty minutes to get the patches on the way down. Using Fedora, about twenty seconds to get the patches on the way down.
Again, one wonders what went wrong here.
Necessity is the plea for every infringement of human freedom. It is the argument of tyrants; it is the creed of slaves.
Do you observe that lately if someone puts Windows instead of Linux is news.
Just like: a dog bites a man is not news, but a man bites a dog is. That's telling.
"It is our choices, Harry, that show what we truly are, far more than our abilities." -- Prof. Dumbledore
If you have an example of an installation that worked right, and are willing to let others talk to you about it, write your success story in a letter and send it to the publishers of TFA.
Do not mock my vision of impractical footwear
Only an absolute moron would admit to that. You have idiots working for you fire them immediately! With absolutely no experience with any unix/linux system and very little windows experience, I setup a mail server, webserver and started creating a website for a company. I did that back in 1996 with RedHat 5 & a Linux for Dummies Book. Linux has come a long way since then. If they can't figure out how to install a modern linux distro in less than 4 hours, you should not be let near any computer ever! I could build a PC clone system from parts and install Fedora Core 4 configure it with apache, mysql, ftp and secure it before lunch. I've done it several times at work.
Finally, an admission of the non-omnipotence of Linux on Slashdot.
But nevertheless, funy how with Windows, the OS is always the problem, breaking the nice, well-written (open source) app, but with Linux, the poorly-written, ill-conceived app breaks (somehow) the impregnable fortress that is *LINUX*.
First, 2-hours vs 2-weeks to install Windows vs RedHat EL3.0?
My first guess is that their own IT guy did the 2-hour Windows install, at which point I quickly look to IBM as the culprit for the extensive Linux install time. Nothing takes 2-weeks to install/patch/whatever. (Even if SAP has specific requirements)
I remember Oracle had a few specifics for the older RedHat 2.1AS. It was annoying, but was manageable. However I would say that knowing the system from basic testing, I never would have used it in production until RedHat EL3.0. The same may be the case for SAP/RedHat between 3.0 and 4.0... Granted the 2-week install was ridiculous, if SAP's Linux pre-requisites require massive configuration time, and SAP and RedHat are certified partners... perhaps they should have done something to jump-start a RedHat install to the right levels for an SAP install. Again I still have doubts about how complex the OS-install is before SAP-install, perhaps someone with hands-on can comment.
Last note, sad to see that there was no saved info from the RedHat crash that could have provided them info post-mortem. Sounds like someone could have acted earlier, since the problem occured every few weeks and they got nowhere after 7 months. Although I have seen vendors ask you to take the system down for 1-2 days of diagnostics... and they may not have wanted to disrupt their testing/config schedule... although in hindsight maybe they should have made the time.
Anyone know if they used the exact same hardware for the replacement Windows system? Would be curious if it was a hardware issue after all and Win32 crashes after a few weeks.
Ever tried browsing the web while running a compile in the background on a single CPU box? Which ran better, Windows or Linux. Do not compare either speed of response or speed of compile, that would be meaningless. Which seemed more stable during the process?
My own experience says that both work well enough, and high CPU & I/O load are not a stability problem in either OS unless some process bumps up its priority too far.
Bob's phone sales in Milwaukee has decided that Nokia is better than Motorola. And Viadork Systems has decided that Toyotas are better than Mazdas. Seriously, though, since when is this news?
Is "Unreliable Linux" a new distribution? I wouldn't use it myself, but it might help keep the consulting fees coming in. Right now its mostly Windows, and I could stand the change. /poor attempt at humor
I love Windows, use it daily at work and home, don't personally need Linux anything - and even I don't believe they're telling the whole truth. Machines don't just "randomly crash". If their computers are doing that - regardless of the OS installed - they have a software or hardware misconfiguration. Just saying "Linux crashed" with no qualifications demonstrates a level of IT competence on par with my mom.
This only highlights incompetent administrators who were unable to get to the root cause of the problem. It's not like Linux isn't stable - hell, even I've had long reliable uptimes with Linux in the past when I used to run it. Throwing out Linux because they have a minor stability problem (that's likely to be a simple fix) seems to be massive overkill, especially when they ignore Red Hat (from the article): "We asked the customer to do a diagnostic test and the customer never responded".
They then go on to complain about Linux auto-updates because they might break applications - but then praise Windows auto-updates in the same breath. Huh?
Confusing? Yes. Are they biased? More than likely.
Sounds like they are having some faulty hardware and not much brains either. I guess any idiot can run some electronics business nowadays.
If there is anything I have learned over the many years I have been a Sysadmin, its that one--with the proper skills, mind you--can do anything he/she chooses with a *nix OS. The possibilities are endless, be it with Solaris, Debian, RH, FBSD, etc. To read about a couple of clowns that don't know how to use *nix, then bitch about how it didn't work for them, makes me sick.
my 2 cents, anyway.
We're like rats, in some experiment! -- George Costanza
Why should we care? How many computers have they deployed? The article has no mention of relevant numbers. Sounds like this is some small, stupid shop that has made misinformed judgements and conclusions about stuff they have no clue about.
Crest Electronics receives two free copies of Windows
I've seen RHEL crash on certain hardware during logrotate -- after swapping out memory, mother board, etc, etc, still the same problem. No idea why, but there do seem to be some versions of their product that don't work right on certain hardware. This was a 3 year old Compaq server that ran just fine for years w/ RH6.something, but I don't remember the details and this contract job, needless to say, ended in embarassing failure.
But on the whole, this is the first such thing I've seen in Linux. Just makes me wonder if this user had the same weird issue.
This really has not alot to do with Linux as I see it. I haven't worked with SAP, but I have with Oracle, I've installed it on both linux and windows systems. The bottom line? I like oracle on linux better overall, but its definately more of a pain to get setup under linux that it is in windows, which I often find bizarre as oracle was primarily written for unix ?? But regardless of your OS, one thing is for certain... there will never be any sort of apparent end to patches, and consultants when it comes to these systems. I'm no programming genius but by what I do see is for what this software delivers they could have made it MUCH MUCH MUCH easier to deploy and maintain. I mean hell quickbooks which takes a whopping 30mins or less to fully setup essentially does everything oracle does, the only real difference is scalability and interoperability. I often chuckle and cry at some of the problems I encounter when working with oracle applications but then it becomes painfully obvious that these apparent mistakes and non-completeness in terms of the software simply serve to generate more money for the oracles, sap's, and great plains of the world thru exhorbent support and consulting fees. Who would pay $50,000 a year for software support if it actually worked? Exactly the point. Its a complete and utter joke to hear these companies talk about how they are gonna make they're software so automated it runs itself. HAHAHAHA... Instead you will get software that automates all the existing admin functions and instead requires people that know how to fix the automated fixer! Which of course will requier you attend the vendors schools to learn how to do, and obviously charge them for the training. The end goal here is to get the level of abstraction so high, that no kid can just learn how to fix it himself, and instead must rely on the vendors total support, consulting people included.
Obviously some asshole incompetent or worse - a shill paid by Microsoft.
Not even worth reading the article.
Richard Steven Hack - This sig is TOO GODDAMN SHORT TO DO ANYTHING USEFUL WITH! MORONS!
I have installed systems for customers of mine that had to be reliable. We have run into our share of problems on both Windows and Linux. However, our experience is that *when used in the right areas* Linux is easier to troubleshoot and maintain. Sometimes braindead applications cause high CPU load meaninglessly, but I have never seen this bring down a Linux system.
Case in point. The most unreliable system I ever installed was a Linux server for a small retail management installation (2 registers, one server that was supposedly fault tolerant, etc). Well, we ran into two problems. The first was that the application was sitting on top of PostgreSQL and was trying to do an outer join between a very large table and two empty tables. Since PostgreSQL assumes that zero-length tables are really say 10 pages in size, it was doing a nested loop join against two empty tables every time the new invoice window would be opened, meaning that the registers were full, the 20 seconds or so was not only slowing the business down but also causing the server to be under very high load most of the time. No problems from that aside from poor performance. We did, however, hack the application to make it stop such braindead behavior.
However, a few weeks after we were able to fix this problem, we started getting database server errors (most commonly corrupted hash table messages). These were intermittant, but usually occurred at about the same time every day if they occurred at all. Turned out to be hardware failure of course.
In neither of these cases were the problems the fault of Linux. One was a correctable hardware failure and the other was a correctable softrware error. OTOH, I am not brave enough to try to run SAP on Linux.
LedgerSMB: Open source Accounting/ERP
One of my personal beliefs is that problems like the ones stated in this article is the primary driving force behind the boom in the appliance market. Small to middle size businesses do not have the IT resources and sometimes no expertise to setup complex applications and maintain them.
For these reasons this market segment is turning to the black box appliance vendors to solve their problems. It is basically outsourcing your IT department. You no longer need to understand the technology and you don't have to worry because in the end you simply tighten the rope around the vendors throat when things don't work.
This is the slow death of the old IT world. Too many wannabe IT Guru's get in to IT shops without knowledge of how to properly maintain and manage complex systems. They may be some of the best minds in the tecnical field but without experience and/or training in systems management they tend to fail as the business grows.
I see it regularly when the company I work for acquires a small business. We review the IT shop and find little to no documentation, no proceedure, and no plan for the future. Just fly by the seat of your pants sweating midnights at the keyboard trying to figure out what went wrong when they applied that patch before testing it.
This story isn't about Linux or Windows, it is about good administrators and bad. Folks need to learn that an Admininistrator is much more than a keyboard jockey. Thoughtfulness, grace and cover your ass back up/out plans rule the day!
--russ
I have seen SAP installations on W2K and RH on identical hardware and boy, the SAP on RH was creeping.
If the SAP folks are neglecting their Linux ports, it's a bad thing that it pulls down the Linux name.
SAP should fully take the heat and blame for not having their shit together on Linux but I have heard they are the "untouchables".
Sad to see Crest brush off Linux this way ... Sorry, wrong Crest. :-) -- Paul
OpenSource.MathCancer.org: open source comp bio
For the uninformed (like me) what does 'SAP' stand for?
Companies are more willing to buy something faulty, but supported, rather than something technologically superior but unsupported. The reasoning is that in both cases, things can still go bad, but with support there's always somebody to take the blame. Windows, no matter how flawed the OS is, is always going to win out here because MS is a huge company with enough money to buy their own small country and support everyone who runs their OS - or at least, this is the image that businesses buy into. It's only in specialty areas where you are going to find customers who are Unix geeks. These people are unfamiliar with the mouse, GUI, razorblades, social skills, and showering, and are going to actively encourage you to hack things - support be damned. Unfortunately this does not represent most customers of IT services. To tell the truth, most Linux installs are a nightmare because of the lack of cooperation between Linux vendors, hardware vendors, support vendors, and software vendors. To make matters worse, most software Linux installs aren't a matter of point and click, and room for error is abundant is such cases.
READY.
PRINT ""+-0
were made for each other.
I find there is tons more documentation about how things work under linux more than other OSs...maybe I just can find it more easily. Plus you can get the ultimate documentation - source code :P
Hello,
I was answering on this :
The admins were probably dumb microsoft certified mouse clickers.
The answer is, like I wrote, in the article :
Crest's IT manager, Anthony Horton, oversaw the deployment of SAP on Linux in November 2004, after inheriting the decision when he took the job. Having previously run SAP on AIX - IBM's version of Unix - Horton was comfortable with deploying such a mission-critical application on Linux.
You claim it could be a HW fault or a sysadmin fault but you forget SAP. Why ? Read the other posts on SAP in tihs thread.
Regards,
Men are born ignorant, not stupid; they are made stupid by education. Bertrand Russel
Mr Horton disagrees: "It might be fine for things like security patches, which don't impact SAP certification rules but with some patches you still actually have to check the release levels and then check against the SAP site. Otherwise SAP might ask you to roll back to the previous version before they will support it." Crest Electronics is trialling Microsoft's Windows Server Update Service
WTF?
How can he say automatic updating is not supported by SAP, then turn around and use WSUS on Windows?
Perhaps he should look at his patch management process and not just blame Linux.
I don't make predictions, and I never will.
When a Linux kernel crashes, it's called a kernel panic. Let me hum a few bars for you:
Sep 26 16:56:45 box1 kernel: Assertion failure in __journal_unfile_buffer() at fs/jbd/transaction.c:1460: "jh->b_jlist < 9"
Sep 26 16:56:45 box1 kernel: ------------[ cut here ]------------
Sep 26 16:56:45 box1 kernel: kernel BUG at <bad filename>:18059!
Sep 26 16:56:45 box1 kernel: invalid operand: 0000 [#1]
At this point, your box is dead in the water. Dead as a blue-screened NT box.
If you've never ever seen one of these in years of being a sysadmin... wow, you've been unbelievably lucky. Buy a lottery ticket RIGHT NOW!
Let me be clear: I'm also highly dubious about the conclusions reached in TFA. The RedHat folks said they asked the client for diagnostics that they never received. There's several other bits that make no sense in there too. But your point - that Linux boxes never crash at the kernel level - is just not so.
I find this a little hard to believe. I implemented the unattended package from sourceforge and the fastest I could get windows to install was something like 30 to 45 minutes. This is not including software and patches. It also took me a week to setup all the bootable media, software repositories, and my learning curve. A custom system would take longer if it was done right. You have to defrag the thing twice after the install, thats another good half hour. I have been running linux on my desktop at home for three years and not reinstalled the OS since the first install. I even screwed up the file system and recovered it. This install is so messed up but it keeps going even through my learning and breaking new things. I would run windows about six months this way and have to reinstall it because it would slow down so much. I run debian unstable and have learned a lot about linux by trial and error and a lot of the mistakes are still evident but the system runs none the less. The only time I ever had linux reboot or hard freeze was because of a faulty network card. Currently the onboard IDE controller is going out linux complains a lot but surprisingly still runs. The only complains I have are it is very picky about burned DVDs and the video may freeze while it tries to read the disk. Eventually it comes back but I can just hop to a terminal and that system still runs. I have also never used a OS that could run even with the system drive completely full. And don't even get me started on the windows swap file.
Who would choose to use a distribution called "Unreliable Linux"?
I love seeing these kinds of stories...You see, I work as an *NIX admin and I've seen my share of bad admins (I've also had to fix their work)...they simply make bad/uneducated decisions. There is really a breed of Windoze admin that has in their brain that since they can install an Exchange or IIS server that they can do anything with a computer and especially *NIX (Solaris, AIX, BSD, Linux, PHP, Perl, Sendmail, Apache, etc)...
They end up causing more problems than they were worth, but they pull the wool over the eyes of whoever hired them and after months/years of bad admin work, they leave the place much worse for wear. This breed of admin has no experience building from source, operating from the command line, or modifying a config file...on top of this their backups are generally poorly done or non-existant...
This is where those of us that know what we are actually doing come in. Our pay goes up and the day is saved. Our employers see the value in a good admin and make comments to others about how well our systems are running...
Everyone and their dog... I don't want a dog to admin my server :)
This sig has been temporarily disconnected or is no longer in service
LOL.. yeah, and many windows servers are used in "heavily loaded" environments.. i.e. Beowulf Clusters, Large Super-computers... etc. Oh, Actually.. Windows can't run in those environments.
=-=-=-=-=-=-=-= - The Celtic - =-=-=-=-=-=-=-=
As I RTFA, I noticed that the finger was pointed at SAP, not Linux or RedHat. They were careful to point out that Linux works fine for them when they are allowed to apply automatic patches. They then point out that SAP forbids this.
Now it seems obvious to me that they've got something configured wrong somewhere and that it's one little stupid setting somewhere. Someone fat fingered something. It happens and it's always a little thing like that what gets ya. That's exactly the kind of mistake that happens when you get into those situations where you manually have to configure lots and lots of stuff.
So, in the end, what they really chose here wasn't an operating system, it was an automatic update system. SAP didn't allow automatic updates in Linux and that was likely the cause of their problems. OTOH, SAP did allow for automatic updates in Windows so there was less room for error (fewer fingers involved).
That's my take on it. And from my point of view, the way this story was posted on Slashdot, the whole thing is one big troll.
. Quit playing Monopoly with Bill. Switch to one of many non-Microsoft products today.
Linux is a certified platform for SAP, it takes weeks,months or even years to roll out SAP (ask ANYONE that has done it). The install on Linux and Windows should take roughly the same amount of time, unless of course the sysadmin only knows one OS and then it's like having a janitor (the windows admin) design the recycling center. Sorry, I just can't imagine how this is even worth posting! Sheesh.. one manager starts a Linux roll-out, another that is anti-Linux takes over the position and claims victory by moving the roll-out to Windows.
=-=-=-=-=-=-=-= - The Celtic - =-=-=-=-=-=-=-=
If i understand this right the problem was that SAP was a PITA to use on Linux, not that linux in itself was hard to install and use. Try setting up for example Hula Mailserver on a linux machine, ten minutes and you have a full fledged mail server with enterprise workload handling and most features to boot, even webmail. I install various apps on Windows also and i can tell that most of them is a real pain to setup compared to the open source counterparts. Documentation on theese systems is a joke. The problem here lies solely on the vendors and not on the makers of the respective OS, be it *nix or Windows.
Most installs on Windows i do is easy on the surface but almost always demands a lot of tinkering to work properly. Installing can take a couple of hours but man, getting that system to work takes weeks. I think this guy isnt a techie and has confused SAP with linux. What he should have done is to throw SAP out the Window, not use Windows to run SAP.
HTTP/1.1 400
...then run test #6 for an hour. The last failures I had initially thought were software related turned out to be hardware related.
Surprise. All that Korean DRAM dumped on the world market may not have been the highest quality afterall.
"Most problems with Linux on the desktop are problems with X."
I couldn't disagree more. There are usability issues, documentation problems, missing features, etc. None of this is caused by X. I have seen _zero_ evidence that X11 is in any way a problem. The protocol is great, and I think we'd be nuts to ditch such a powerful, network transparent facility. As a developer, I'm not fond of the Xlib APIs, but there's work to replace Xlib now. The XFree86/XOrg implementation of the server could be better built so that it was in many small parts - but that's only a problem for people doing lots of low-level distro hacking, and for distributors. Again, there's progress to modularize it anyway.
X11 is not slow. Some X11 drivers are slow, but that's a driver issue and changing the window system will still leave you with crap drivers. For that, you need people who really understand the guts of the hardware, and you need good documentation. I should note that my system is *extremely* snappy under X11. In general, I find decent ATi and NVidia cards get very good results. If you're talking about 3D, that's in my view quite separate - but again, comes down to driver support and no documentation from vendors.
Nothing in X11 makes apps that use X11 ugly. Seriously. It's *WAY* too low level. Your complaint is most likely with the toolkits, themes, etc. If not, I'd be interested to know what in X11 you think causes the problem.
I'll certainly give you the points on X11 configuration and maintainance. I personally find it pretty painless, but then I have good hardware. I also find X11 to be very stable, though there have been times in the past I've sworn rather loudly about it (usually due to bad drivers or hardware).
The VT system could work a lot better, and I'm looking forward with enthusiasm to the move of much of the frame-buffer programming back to the kernel where it belongs. That should help solve a number of irritations.
I suspect you may have hit the reliability nail on the head if you're talking about rebuilding Xorg/XFree86, fontconfig, etc. If not done very carefully and with a good knowledge of the system, you'll quite possibly break things here. In particular, you need to be 100% sure that your new versions are ABI-compatible, unless you isolate them and only use apps you built against them with them. Your comment suggests that you do not, since Fontconfig has nothing to do with font rendering, and if there's anything you should be rebuilding (but you don't actually generally need to) it's freetype.
Of course, I find I get extremely good quality fonts anyway, so I can't say I've ever felt the need. Fonts under Linux used to be horrific - eye searing examples of pure horror. This has, in my view, been entirely resolved by recent freetype libraries and the ditching of X core fonts in favour of client-side rendering.
I personally find X11 one of the most attractive things about Linux. There are some issues with the implementation, but the power and flexibility of the protocol is not something I'd want to give up. I do agree that it could use some more work, but I'm unwilling to whine about it when I lack the time, skills, and motivation to do it myself. I personally think the current X work is important, and it looks like it'll lead toward more radical enhancements once the more basic issues with the codebase are addressed.
In many shops, the lowest TCO comes from the setup your IT people know best.
"Good" generic TCO studies will assume greenfield conditions, where the IT people know nothing about computers beyond what is common to both systems, or at least that IT is equally-skilled on comparable systems. Unfortunately, "equally skilled" is an elusive term, leaving room for manipulation by a biased study-creator.
The only real TCO study is the one that's custom-tailored to your particular situation. Unfortuntely, unless lots of money is at stake, the cost of doing a thourough study will likely outweigh the benefits. The more "equally skilled" your people are, the more likely this will be the case.
The bottom line:
Most IT shops are skill-heavy in one system and skill-light in another. Unless one of your goals is to balance the skill set, sticking with what you know will save a bundle in training costs.
Knowledge is how to play a game, intelligence is how to win, wisdom is knowing what game to play.
This event is good information, if not good news. Take it in stride.
RHEL3 isn't really linux. It's a dead end frankenstein kernel with more bugs than... well than windows, and redhat ignores them. The only way to run EL3 is on a real linux kernel. The 2.5ish backported 2.6 2.4 kernel in RHEL3 is just stupid. My IBM hardware locked up all of the time on it too. Neither RedHat nor IBM could find the problem. Going to real linux kernels fixes the problem. There is definitely something horked in the experimental IO subsystem they yanked out of 2.5. RedHat unfortunately has a long history of boldly pushing experimental code into their kernels then forcing their customers to live with it after it is proved unstable. EL4 is a bit better, but red hat is losing ground. Shit, their sales people can't even be bothered to send a quote for bulk pricing.
To say that because you've never had problems, the whole "Windows crashes is FUD," is really quite arrogant. Windows certainly crashes a lot, although more often on the client side, where the user can do more damage. There is also the problem of diagnosing crashes once they happen. I've found it much easier to diagnose *nix crashes than the infamous BSofD. There's also a question of motive. M$ has millions, perhaps billions to lose if they don't sell. There's little profit to be made spreading FUD about Windows, while proprietary software companies do have a lot of money to lose to *nix users. Even companies that make money from linux are always vulnerable to a customer switching to another distribution. If you follow the money, it doesn't lead back to linux. FUD is pretty specifically a corporate strategy.
http://en.wikipedia.org/wiki/Signature_bloc
We have PeopleSoft with UDB and it is horrendous. At least an extra year in implementation. I'd say go Oracle, but with Oracle in the apps business, its probably broken thanks to SAP.
No wonder they were forced into Windows... what would be your choice? Linux and UDB? Wow... no thanks.
I *know* this guy is not an admin. He is MIS, at best. *Huge* difference. This guy gets paid to write reports and macros for applications for whatever software this businesses uses, clearly SAP, not to install or administer servers.
I mean, just listen to him. He outsources everything. He seriously believes all operating systems are the same. He complains about having to spend two days a month updating and testing. Then he goes on to include this work in an increased "total cost of ownership" for Linux, completely ignoring the fact that it's his job and he's being paid whether he does it or not. He doesn't know the difference between an application failure (core dump) and an OS failure (panic/oops). And, to top it all off, he thinks autopatching is a great feature.
Lots of "small" (multi-million dollar) businesses make the mistake this one has: they think they can get away with having just one "admin" who is really MIS, who spends all of his time dealing with the business side of things rather than the computing side. To maintain the illusion that this is a workable combination, they switch everything to Windows and spend almost as much on licensing and consultants as they would on a competent admin. Then they wonder why their customers' credit card numbers mysteriously show up on the 'net.
News flash to all the "small" businesses out there: well-maintained computer infrastructure can replace 50% of your employees. Skimping on IT personnel is a stupid, stupid mistake. You can afford to have *both* a proper IT guy and a report-writing business grad. Despite their misleading marketing, Microsoft software is not a substitute for a qualified admin.
"I assumed blithely that there were no elves out there in the darkness"
While I agree with you Linux definately not good for any type of ERP, SAP yes is a great windows product but it was build and designed around windows. However there are a great many others out there, QAD for one example again will make a Linux box puke, but runs like a dream on something like HP-UX. So many linux people seem to forget Unix is still around and still kicking strong. Unfortunately until someone specifically targets Linux for an ERP system it's a poor choice for that type of application. Way too much stuff going on in a ERP system to really get anything to just port over it requires a lot of code rewriting.
I mean think about it, every ERP app I have ever seen, has accountants hammering away crunching financial numbers, warehouses updating inventory, shop floor systems feeding data in, fork lift drivers with wireless devices moving inventory around loading trucks. EDI data flowing in and out to suppliers, customers, banks. Scheduling people scheduling machines, payroll people paying employees, all this on one system. Also take any fortune 500 company and they are doing this in several locations, several countries, several currencies, several timezones. Linux just isn't there to that level yet.
No ERP system will run well on MySql either It just isn't that robust of a DB for this type of application. Oracle or Sql Server are the primary targets in those applications.
We run 10 server with 2.4.x and 2.6.x kernels, with loads in 1.7 - 4.3 range during most of the days. The servers run non-standard communications servers. They are perfectly stable. They were even stable under loads of 10.0 - although perfomance sucked.
I'm a long term linux user.
I have definitely encountered dependency hell.
I started with linux 8 years ago and setup a slackware box. I didn't know anything, and ended up fubarring the system after about a year of installing new software on a regular basis. I tried Red Hat for a while 6+ years ago, and it ended in complete dependency hell. I despise RPM's to this day, despite any software progress, my hate still burns.
I tried Debian for several years, and I had dependency problems from time to time, but never had system meltdowns from it.
I've been on gentoo for a couple years now, and had some issues, but far less than even the debian install.
Basically, as each version of distro adapts and gets more mature, the dependency issues have lessened. You could say that its because I've learned a lot about proper administration, or you could say its the software. I tend to think the various distros are learning better ways to resolve issues.
I must second this, most Linux distros are actually easier to install than Windows and that's saying a lot. These days a corpse with post-mortem reflex could install Linux given that their hand clutched a mouse.
If it really took these people more than a day (I'm being quite generous here) to install Redhat then there must be something severely wrong with these people.
I'm very suspicious of this story.
It's true no man is an island, but if you take a bunch of dead guys and tie 'em together, they make a good raft.
I made a similar realization when I noticed a banner ad for "News Blog", that pointed to a major print newspaper... on a blog site.
"I assumed blithely that there were no elves out there in the darkness"
I was the IT Manager at Crest Electronics many many years ago, not the one that forced this guy to inherit the 'SAP' decision, and I can't really comment on the current infrastructure, but I can understand the environment he works in, and don't envy him.
There's always more to a story though than what's in the article. Creat originally ran on an old DOS based Novell network. In the time I oversaw the migration to a Unix based accounting system and it was fairly robust, and did the job. Like all accounting systems, it had some issues, but its reporting was good and that did the job.
The guy who they employed to replace me, they employed the week before I left, despite giving them 5 months (yes 5 months) notice of my intention to leave and do other things. He was pretty keen on introducing new databases right from the start, so its no surprise he went down this track eventually.
Anyway, the current guy they have they actually employed him before they even told the previous guy to leave (so I heard), which is kind of pitiful, but that's the way the owners can be. I guess they really wanted him gone.
It probably wasn't really Linux, SAP, Redhats fault, it was most likely sloppy implementation, coupled with a business that expects a lot for very little. (Whoops, did I say that?)
I smell a useless windows admin tasked with running a linux server(s). I deal with shit like this all the time.
It sounds like you think Google buys a few hundred thousand computers, loads up the latest kernel, and just starts running! If they did that, they would probably stay up for less than a minute. Google is not running IBM hardware or RH Enterprise (and obviously not SAP).
No, Google uses custom hardware, custom kernels, custom drivers, and custom everything else (web servers, filesystems, etc). They have gone through immense pain fixing unreliability at heavy loads for all of those things. You might think that ext2 or a driver for your favorite hardware is rock solid, but that's only because you haven't run Google on it.
The only way Google gets scalability from Linux is to use lots of cheap hardware; each computer has a single CPU, cheap RAM and a couple hard drives. Each machine does the same job as dozens of other machines, so that when one falls over it is hardly missed. Google can afford to do all of this because that is their job. Crest's job is selling electronics, not supporting their ERP system.
Of course all of this customization is possible because they have the source to all of the software they run. On the other hand, I suspect that MS would give them a source license if they had 100,000 Windows servers. FYI, I believe that MSN has on the order of 30,000 Windows servers.
dom
http://packages.debian.org/stable/devel/crashme
Do you seriously think any Windows would survive weeks of that kind of abuse? Hell, you'd probably come up with a new virus that would immediately send to everybody in your address book, then crash the system.
"City hall" in German is "Rathaus" Kinda explains a few things......
I'll readily concede X is slow and ugly, but if its unreliable for you, its because you're a retard who should stay far away from linux systems. The only time I've seen unreliability in X is when one is using proprietary drivers from ATI or NVIDIA, in which case, they're only a little less reliable than the Windoze machines. I have never seen X crash with normal desktop use. When lockups occur, its alway because of the application doing something its not supposed to do. I don't run the application, and golly gee, I never get a crash.
Yeah, if the extent of your computing requirements is to play a computer game with voxels, you shouldn't be using X or linux.
There is no America. There is no democracy. There is only IBM and AT&T and DuPont, Dow, General Electric, and Exxon
I was with a company that wanted to go with Coldfusion MX so we rolled it out on Linux. The system would constantly crash and the support forums for it was full of complaints... for the Linux version. It ran fine in Windows, but the company couldn't get it to work right under Linux. So we went back to PHP and never saw a crash again.
The problem wasn't Linux, it was that the product was written like crap.
But I imagine it's easier to get crap to run under Windows than Linux, because Windows has less variance.
Looks like someone should have hired better consultants, or realized it before they implemented when their scope was defined.
It's entirely possible, they had complete idiots implementing their software.. OR, that the software itself was horrible.
I would also guess, they didn't know how to troubleshoot their own problems.
Everyone knows the Apple Store, one of the largest online stores, runs on.. oh, wait.
We do know that Macs are useless for clustering and could never be used to build a supercomputer.
I know, old ideas die hard.
Read the EFF's Fair Use FAQ
This thread seems especially pointless, even for /. -- there is not enough information to infer anything at all. There is no such product as "SAP", so no one could have installed "SAP" on Win32, Linux or any other OS. There is, however, an SAP AG that is a vendor of an imense suite of ERP products.
I am currently involved in an SAP Netweaver '04 implementation at one of the largest SAP customers in the SF Bay area. I have to admit that I have no experience with SAP software on a Linux platform -- my experience is with ERP 2004/Netweaver 2004 on Wintel and Solaris. Even so, I think I am accurate in stating that any significant part of the suite that you install on either of these platforms would not be useful in just a few hours. You probably won't have finished installing the base components, the patches, the service packs and the relevant business packages until towards the end of the first day. And then you still would not have even begun the lengthy task of configuring all the backend architecture to play together. And keep in mind, this is NOT a single server business solution, even for the smallest SMB customer!
So, what exactly does it mean if someone claims to have "SAP running" on a box in a couple of hours? It sounds kind of like a mail server with no network interface -- runs like a champ for months on end, no problems!
Maybe I'm missing some deeper insight, but this so-called "news" tells me nothing about SAP, Linux, Windows or Crest Electronics. Nada. Zilch. Click the back button and keep scrolling folks, there's nothing to see here.
I live in Brisbane, and the area where Crest is situated is renowned for power supply problems - only the best UPS's will help. I'll bet that the guys who "fixed" the problems with WINDOWS supplied a new UPS with their gear.
"We got to the point where we had a business requirement to move on, we couldn't wait for the error to occur again, because it was not a reproducible error,"
So it seems the error was not so frequent, nor reproducible?
I'd say this probably was some hardware issue, if this was real at all. Sounds like Microsoft marketing/PR people wrote it. It drives the point through (lower TCO, instability(eh?), Windows-is-better) while at the same time doing so relatively stealthily.
I also deal with this shit all the time. Find me a MCSE that understands any real OS and I'll show you a salesman.
Unix? What's that? Microsoft invented the IntarWeb! Ack.. die you stupid fuckers.
I know some of the people involved. I doubt that was the reason why, considering the decision was taken after OS X was available.
HBI's Law: Frequency of calling others Nazis is directly correlated with the likelihood of the accuser being Communist.
They are using RHEL3 maybe they had one of those problems with kswapd. We've had problems with kswapd on RH9 too. Had to reboot a server every few days. So yeah, things aren't all that great despite the great faith some fanatics have in Linux.
If you bother to look, Linux isn't quite as stable as some people believe it is, at least for some versions of Red Hat Linux. Makes you wonder what Red Hat are doing. They are supposed to be making their kernels more stable than the developer kernels (which aren't that stable - the kernel developers nowadays don't seem to care as much about that).
This article is nothing but FUD commissioned or created by Microsoft - not to mention that it's full of lies. I have run Linux for years and my total cost of ownership has been $0.00 and I haven't had a single problem. This article is typical of the lies Microsoft spreads because it is scared of Linux taking over - which Linux is doing quite well.
when you have your servers up in 2 weeks instead of two might cost some dough .... but stability costs less at the end ...
...
...
.. if you have 24hr support sitting on a reset button windows might be OK, if a reboot costs you heavy dollars and long distance calls and several minutes of services down you should choose 2 weeks install and no reboots..
when your servers on windows will blue screen at the middle of the production day that wiwill most likely cost a lot more on the long term in productivity loss and people sitting in their offices not being able to access resources
yes i can install windows box in 30 minutes with webserver, however i have bsd boxes running 365days+ with dns/apache restart and having a good sleep while my non windows machines run is just cheaper me than having a blue screened server for 8 hours and loose customers or receive pages to "fix that crap" in the middle of the night
of course your mileage might vary
just a note: how can an installation of a software last 2 weeks vs 2days ? Same software ? I know sometimes clicking a defult config together takes less time than building a config file (text) from a bad template/example but 2 weeks ?
God created all that in 1 week! (including basics for SAP and Linux in a way) -OK He knows more than us I guess
RTFM N00B!
Get your Unix fortune now!
this is a server we're talking about, it's not supposed to have a gui ;P
sum.zero
they're just down right lazy....build a custom kernel for your system....period.
A stock install is not for production use anyway....just like a stock Windows is.
if you dont know what,how or why someone would use a custom kernel you have no damm business using it in you hardware.
windows people know there software thats why they swear by it.... as we *NIX's do....but as a company that is DEPENDENT on an outside OS like LINUX...you must ONLY care about selling UNITS like the Music Industry trying to sell crap music.
that we are going to find out that the baby's daddy is really PeopleSoft?
I mod everyone down who says "I'll get modded down for this." I hate to disappoint.
....coincidence? I think not.
Now wash your hands.
Maybe they should try a distro other than Gentoo.
Karma: It's all a bunch of tree-huggin' hippy crap!
Yes, I did read the article, and I actually use Gentoo myself and love it. :-)
Karma: It's all a bunch of tree-huggin' hippy crap!
... who takes two weeks to install a piece of software is absolutely incompetent.
In such a case, it's certainly much easier to blame the operating system than admitting it...
The Tlog - a technology blog
Someone could easily say 'I run SuSE 10' and that would be the same thing as saying I run FreeBSD (in your terms of knowing where things are). I totally see your point, and I believe that in the mid term so many distros is killing linux, and spreading so many developers so thinly mean so much work gets done so many more times.
When people say 'choose the distro that works/suites you' I call bullshit. Most distros are the same in terms of packages available, and if they are not, this is merely a triviality of choice, all packages could be the same in all distros, so lets strike that off the list.
Which desktop environment/windowing environment does it favour? Strike that off the list - all distros could support both, or any one.
So a distro is a set of configuration and package management tools, some additional distro proprietary configuration of default settings and drivers and kernel patches - plus lots of other tweaks and stuff.
How does this imply that any one distro is better at any type of task than another?
Get one set of good package management tools, is apt-get better than emerge? rpm's better than Linspires on click installers? I say that any application that can be packaged as an RPM can be released in any other format, via any other distribution means with some small work. So, standardise on package management would be one good way to help alleviate this issue.
Instead of having an install base for a certain configuration of kernal patches in the hundreds, why not have fewer CERTIFIED (by certified I mean they are named) kernel releases and configurations that are then tested on ALL linux user machines. Then each kernel convolution would have more testing and users would be less inclined to jump ship silently until they go through 4/5 live distros to find one that boots cleanly, nicely, and then install that one.
FreeBSD is the name of the FreeBSD kernel and the FreeBSD operating system. Linux is the name of a kernel, and used to describe operating systems of all varieties built on that kernel, be they gnu/linux or ftoo/linux.
I use Linux, is as you say, an stupid thing to say. What car do you drive? Oz Racing locking alloy wheel nuts. What? Yeah, thats my car. No thats you wheel nuts... Yeah, but you know, that is how I describe my car. How many doors does it have? 3. OK, what colours is it...
In the longer term the opening up of so many distros has expanded the number of teams working in different ways on certain features, which will improve these areas, and perhaps this is why linux is seeing more public adoption that FreeBSD. Saying that I would never choose linux over FreeBSD for a webserver because of the simple fact that my last 3 web servers have been FreeBSD and I like to know where I stand, on what should be, a utility computing source. A webserver, which runs basic http protocol SHOULD JUST WORK in 2005. (and all the garbage vm's and libraries that power the sites).
People don't generally say they run XNU or Mach, they call the operating system by another name in that case, or the hybrid kernal of the Mac. And people do so in linuxland - I run gentoo, I run yellowdog, slackware, debian... very important distros. And perhaps these distros ARE the 'certified' configurations that will get more testing, but I still see them as too numerous in the short term for linux health (think companies wanting to run SAP on their linux boxes, the certification and support hell).
Thats all folks, any errors or emissions are intentional.
please type the word in this image: bylines
random letters - if you are visually impaired, please email us at pater@slashdot.org
#hostfile 0.0.0.0 primidi.com 0.0.0.0 www.primidi.com 0.0.0.0 radio.weblogs.com
I am happy because I managed to install linux only in a month and I did it from ***SCRATCH*** :)
I wonder why it took them so long to install it!
sex is better than war!
thats says it all in a nutshell i think. he's a retard.
If you mod me down, I will become more powerful than you can imagine....
If you have an MCSE certification it doesn't automatically make you competent to administer a Windows network.
I've had MCSEs call on me for help with simple networking problems.
I find that many qualified people just forget what they've learned. I even have the same people calling me up every once in a while, with the same questions, purely because they keep making the same mistakes.
It may just be coincidence, but, I find that the most incompetent MCSEs are those who go out of their way to tell you they're an MCSE. They seem to use it as an excuse for their incompetence - like saying "Well, I was smart once!"
: )
Linux/Open Source/Anti Microsoft News
Frigging Red Hat. Gives Linux a bad name. I feel for this guy, I really do. Installing and getting RedHat up to date is a major pain. Gentoo and Debian don't have that problem, but they aren't marketed like Red Hat, so the uninitiated have no idea what they're missing.
Stop-Prism.org: Opt Out of Surveillance
Is it just me, or does this sound like an ad?
"I fully support Linux but if I had to make the decision again I'd pick Windows. A big reason is the fact Windows was up and running in two hours at all the right patch levels. The installation of SAP took two days on Windows, the installation on Linux Red Hat took two weeks. The total cost of ownership is actually lower in this case than with Linux because of the hidden costs of the support."
I feel like I'm reading a Microsoft brochure. And keep in mind that I *like* Windows as a desktop OS, for the most part.
A good Linux Sys Admin. Just one was enough! He'd have adone all the work for you in --Evaluating HW/SW --Drivers -- Compatibility --Security And please don't tell me Windows is better option, Sure if you don't know what you *really* want from your deployment, then go for windows, there it fits perfectly fine. Just click next and all is done.NOT!... Look at all the *openness* that makes it a ready target to all kinds of breaches (So is Linux BUT it can be secured with more intelligence than on Windows level). Anyways Linux just needs a Good backing. Novell is beginning to do a good job on SuSe/Open Exchage Servers etc.
Scott McNealy to Michael: "Suck my Sun!" Michael Dell to Scott : "Lick my Dell!"
So in other words, hardware fault that they never bothered to trace. What's the bet the Windows system was on fresh hardware?
Someone hates these cans.
This story is only possible if the chap in question has a chromosome too many!
ROFL!!!
LMFAO!!!!
PIMPFL!!!!
-if at first you don't succeed, stay the heck away from paragliding.
If IT-managers can't get prober specialists to do the job,
is changing the software a solution?
I wonder why they didn't use Suse Linux. Suse and SAP are both German and they have worked together elsewhere. SAP is supported on Suse and for this purpse I would think it would be better than RHEL.
See my journal, I write things there
I run a 2-core HT Xeon system @ work, it looks like a 4-way CPU. This is great for finding race condition bugs in my software -and other peoples.
my core troublespot appears to be cups, which will spin at 100% of a CPU within a few hours of starting. So I have to restart cups every morning. This is so, irritating. I suppose i could just get cron to reset it for me, but still.
Whereas on windows, I havent had to reboot my laptop since, what, yesterday, when the clipboard stopped working. I didnt even know the clipboard could stop working, but no, you can suddenly stop being able to cut and paste. Trust me to be the one to find out.
I use "Unreliable Linux" all the time it may not be the best but it's free... how many other distros can say that eh, mr clever clogs?
I've installed SAP r/3 on Solaris and Linux... never a problem. Yes it's complex, but hey - that's what's fun about being a sysadmin! One thing to keep in mind folks is that it's a GIANT fucking database - tons and tons of tables. And like all databases, you have to care for it and feed it.
That means watching memory usage, extents, indices, disk controller utilization, network interface utilization, swap space, processor load, and on and on and on.... And trend it all out... then monitor some more - when you see the problem, you fix it. Sometimes that leads to other problems that get exposed as you move up the line - but when it's all done the system rocks...
I know this because I've built HUGE systems used by thousands of agents every day... they run on Sun E10000 boxes, E4000's, HP K-class boxes and so on... If you're installing SAP and you can't handle the diagnostics - get the fuck out of the computer room and go back to asking if you'd like fries with that...
I read the article and thought that the guy had a bunch of dingos for admins...
Linux had its prime but it's mostly over, it sure won't be on a FREEBSD status but will never aspire to the top again. Thanks to the so called linux consultants that wanted to cash hard and fast, your greed killed Linux, thanks for that, well, drink a beer and forget this issue at all.
One should remember that RedHat (and Fedora) is one of the distro vendors who add the largest number of patches to the linux kernel. And now that RHEL is so much separated from Fedora, their particular combinations of patches is probably far less broadly tested. At least under several occasions when investigating severe problems on RHEL boxes, the culprit appeared to be a problem with the kernel, which did not appear in a same-generation Fedora kernel...
So if their team was so motivated/competent with linux that they required 1 week to install a RH distro, it is no wonder they did not investigate much. Did they even call RH support to get the problem investigated ?
What is this, slow news day? Windows NT/2000/XP/etc is unstable on flaky or misconfigured hardware, and stable when done right. I very much doubt that Linux is any different.
My Karma: ran over your Dogma
StrawberryFrog
* Disclaimer : I used to work for a competitor to SAP, but don't anymore. I still find this absolutely hilarious *
This is a couple of blog postings that one person's experience with installing SAP. They're long writeups, but pretty darn amusing (unless it's happened to you of course!).
Part 1
Part 2
Part 3
Part 4
You really can't make this sort of stuff up.
when people, who mostly use windows and which have at least a little base knowledge of linux try to do a professional install of linux, on a machine not suitable for production ... let me guess
..
nforce chipset? promise raid? you wonder why its not stable? or what hardware setup did you really use?
of course, if you want a stable linux system, you have to look for the hardware which has stable drivers in linux, but thats too much thinking for a windows pro having a little linux knowledge
those crashes wouldnt have happened with admins who know the hardware they order!
and installing linux including patches is possible under 2 hours!
Wow, over 500 comments, mine included, on an article that was checked for technical accuracy based on a weak source with no supporting logic or evidence to support the final conclusion.
Amazing the power of the Internet, all these intelligent minds, effort and time to combat such a weak article and source.
I guess slashdot ad revenues must be down again.
"No, it's efficiency and good systems management."
Not for servers its not. You never "auto-patch" any server, much less a production server. If you don't understand why, they you've never been involved with operations or supported a mission critical application.
And he isn't the least, tiniest bit a paid shill to spread patented MS-FUD!
No, sirree. But you never saw a better recipe for a flame-war on Slashdot. How should we do this one? Doom-style? NeverWinter-style? Quake3Arena-deathmatch?
I meant, of course "Linux to Windows". Guess I'm too used to typing it the other way round...
FreeBSD in a way is just a distro of BSD, as BSD in OpenBSD, FreeBSD, NetBSD, all of which swap code.
Saying FreeBSD is not fragmented like Linux compares two different things.
FreeBSD simply lacks the distribution even to be fragmented. It is a very cool OS for that, but an argument advocating FreeBSD over Linux will inevitably end up saying "and FreeBSD can run linux binaries too!".
Retract the previous correction...I worded it right. I *hate* head colds, they screw up my concentration (the rest of the time, I have some other excuse)
SAP is a piece of software that is perfect for you if you enjoy pain. The interface alone will give you nightmares.
"A great democracy must be progressive or it will soon cease to be a great democracy." --Theodore Roosevelt
If any of the drivers you are using are reverse engineered or not well maintained then you will get problems when running under load (ie. on a server).
You have to pick the hardware to suit Linux, not everything is that well supported, some vendors don't like to see their register layouts exposed for some reason (even though you can probably find that stuff in datasheets).
I think you meant "... are rarely used without modifications and customization
This sig under construction. Please check back later.
for the link. Otherwise, you're a total asshole, but it's nice to know that you aren't a total waste of carbon atoms.
HBI's Law: Frequency of calling others Nazis is directly correlated with the likelihood of the accuser being Communist.
it was done by an incompetent idiot. Sheesh, installing Gentoo shouldn't take that long. :-)
The funny thing is other people have told me the same contradictory stories about SAP, so I half suspect the in-house developments.
No contradictions about the cost tho.
Do some compatibility research (even if it is just with Google) before buying expensive hardware.
This is a problem I have noticed with Linux many times. On the whole, Linux is incredibly rock solid. But there are rare instances where specific combinations of Linux software and hardware will cause crazy problems. For example, at one time there was a problem with APIC in the kernel. If you had an nforce2 motherboard and a kernel with APIC enabled it would freeze up semi-randomly. 99% of people did not have this combination, so it wasn't a problem for them. But for the 1% of people who did, how were they supposed to figure it out? Only if they are very involved in a Linux community would they discover this.
Another problem I had was with the combination of Ubuntu, Nforce2 IGP, the NVidia driver and not having DDR Dual-Channel enabled. This combination brought about frequent freezing. But who could know without good googling skills that this combination was the cause of the freezing?
I'm willing to bet that this guy had one of these weird combinatory problems. It just goes to show that the Linux testing procedure is not 100%. But switching to Windows when this happens is basically just claiming ignorance instead of figuring out why it's crashing and fixing it.
The GeekNights podcast is going strong. Listen!
No matter WHAT system you run, if the tech jocks give you the run-around - you call in independents - just like getting a second opinion - Doctor, Lawyer, you name it.
Start by posting on forums. Free answers, or find out if anyone else has/had the SAME setup. Then dump/burn the key configuration data for professional third party opinion, and burn that snapshot. That puts the heat on consultants, because if an IT manager does not have basic problem solving skills, they SHOULD.
Post windows, examine and compare the parmlib/config/ settings - any differences?? The one thing about SAP, is that their consultants/partners/associates are not supermen. Fish out those backup tapes, and get ready to 'post' them. Big fees carry professional negligence, or you can point to failed 'service level agreements'.
So said, start disaster planning planning your next windows/worm/virus/Service Pack strike, because 2 days is not enough to get a handle on this. Best of luck, and do post how things turn out 12 months later.
2 weeks to install Red Hat?? !! I don't belive it!!
Then I guess when I took my RHCE (Red Hat Certified Engineer for those who don't know) exam, the two hours I was given to (successfully) install a Linux server from scratch (including a kernel compile & installing/configuring a few network services) was all just a dream, was it?
I suspect their "installation engineer" probably had little Linux experience in the first place.
Gentoo Linux - another day, another USE flag.
... when I need them? That is so true. I expect Windows to catch up in terms of security and stability with Linux/Unix but to lose out in usability/userfriendlyness as Microsoft tightens the bolts on their OS. So while Windows will definitely catch up to Linux/Unix in terms of being a server system that should be taken seriously, expect the bar also to rise dramatically on the demands for knowledge, education and professionalism being made on Windows administrators. Alot Windows Server operators are going to find them selves looking for more highly qualified computer professionals because their old self taught went-to-night-school-and-got-a-certificate Administrators can no longer solve problems thrown up on Win 2003 or Vista Server that weren't there before because Microsoft left all hatches open to make things easyer for these less well educated admins. I already see this happening and the amusing thing is that after installing Windows 2003 many of these Certificate Admins solve the problems created when Microsoft changed default settings in Windows to a more secure configuration by reversing those changes and opening the security holes up again. I suppose you could argue these Admins are just being lazy but I think that a simple inability to adapt to the new more secure way of doing things has just as much to do with it.
Only to idiots, are orders laws.
-- Henning von Tresckow
...the summery going... summary
...death sentance. sentence
...an implimentation on... implementation
...that doesnt fit... doesn't
I give this service for free...
They didn't do any configuration testing? Take typical hardware and run a bench top install to see if there were going to be any problems? And the engineer asking the customer to run a diagnostic, wtf? Get your fanny in there and nail down the problem, don't ask the customer to do your support. That should apply in *nix as well as Windows environment.
That's our life, the big wheel of shit. - The Fat Man, Blue Tango Salvage
I had a few GPFs in Linux once back in the 90's
The computer was find after that for the next 3-4 years before I threw it out, didn't seem to be a hardware problem.
IIRC it was 1.2.13, and I didn't set up odd settings.
These things do happen, they are just pretty rare.
And why would anyone run SAP on Linux?
My current current uptimes:
Linux Redhat ES 2.1 = 330 days
Mandrake 9.2 = 192 days, 245 days
Windows Server 2000 = 23 days, 40 days, 33 days, 70 days
Although Linux may not be the right choice for every job, it certainly seems that way to me.
Bollocks. Linux does not blue screen and core dump at all without a reason, and it is impossible to find out whether it was the OS itself or the pile of crap called SAP he's running. His information on this is just too vague for any sort of IT person.
This is a guy who decided he didn't like Linux when he'd started, and that goes for his feeble attempts to solve the issue. Another clue is he then goes whinging to the press about it. If you don't do what the support people ask then they cannot help you, and I daresay you'll have the same issues on Windows as well at some point. And no. You cannot get a fully working Windows system up and running with SAP in two hours, which is a complete lie. There's the setup and testing as well which will take the same amount of time.
This guy is someone who thinks that installation and deployment is clicking Next -> Finish in an install wizard. He couldn't piss up against a wall if he could find one, and quite frankly, I don't consider incompetent arseholes in IT to be particularly newsworthy. The fact that most of them happen to choose Windows is a bit of a clue.
Crest Electronics IT department deemed incompentent, laughing stock of industry.
The difference between Canada and the USA is that in Canada healthcare is a right and gun ownership is a privilege.
I understand the love for Linux (or rather hate of MS). To each his own. However, in this situation it is a common problem that companies are having with Linux. 2 weeks to configure and deploy. Come on, that's ridiculous. This guy had a business need and needed to get the job done. If you can't understand that, then you'll be unemployed 6 months after convincing management to make the switch to linux and you still haven't gotten the entire enterprise up and running. Also, and read this slowly so that you don't pass out: LINUX CRASHED.. Now, the hardware was blamed and then the administrator. What's next, code fairies???
Nobody is saying that today, so don't let people on with rubbish about that someone might say that with your "anyone that says" and "in every situation"-crap. Nobody says Microsoft Visio runs best on GNU/Linux.
If you have chosen the product that best suits your needs, you need the platform it runs on. I don't know a single business that runs only one piece of software. So in addition to asking people to choose the product that best suits their needs, you pretend that any piece of software is available on any platform and that all the other pieces of software they need also are available on any platform. Following you, it would be sensible if every app came with it's own operating system.
I'm happy for the firm in question that Microsoft's and SAP products was easier and more stable. However, there is no magic here, just some bad engineering somewhere. A GNU/Linux system with no logging? Just died with no trace anywhere? I find that hard to believe.
I read TFA and read it again, no test system was mentioned. My SAP days are over, but I distinctly remember that having a test system was a requirement, means there should be a comparable system to test patches and customizations to SAP and the OS first. If everything runs OK on the test system, then you are allowed to apply the modification to the production system. /etc/syslog.conf tells you a lot about your running system. You have to understand how to do it, though.
It seems that that didn't happen and that the SE(s) in question had no clue how to handle the installation in the first place. Two weeks, give me a break, none of them did an install for a production system before, I guess. Repeat with me: It is not a good idea to learn on a customer's system!
A little hint for the poor soul who did the install: a well-groomed
just my 3 cents
I guess the story wrote itself. "Anything is supposed to be more reliable than Windows, but these people tried Linux and the decision blew up in their faces and they had to crawl back to Microsoft to get a reliable operating system!" Then you read it, and it isn't even about Linux, it's about some bug in a proprietary software package. The author of the story gets credit for spinning it! Why do people read the computer industry trade press again? Oh, and "the hard work required to keep Linux up and running" - huh? I often forget how I set stuff up on my Linux servers because they run for months and years without me touching them. I have to make notes so I don't forget what I did. I don't suppose anyone has ever factored in productivity into TCO - your admin guy is solving interesting problems on a stable Linux box, while your MSCE is patching MS' mistakes every month.
I had the very same problems with Red Hat Linux. So I went back to Slackware.
now we need to go OSS in diesel cars
I don't see how you can have a system core dumping without a major unaddressed hardware failure. I run several production servers on the bleeding edge with regards to kernel level and application level. The only failures that I've ever experienced on any of our servers is hard drive failures. So, I'm thinking this was paid for propeganda in order to falsify reasons to which this change was made.
I'm a subscriber to the newspaper quoted, and my experience is that a lot of the "articles" in the NEXT IT Section are fairly obvious re-hashes of press releases, media handouts, etc. A quick Google for the article's author shows that he was Deputy-Editor of the paper's IT section until April this year, when he resigned and went freelance. God knows if he was still working at the paper when this article was written, or whether it was written after he left. But I'm enough of an old cynic to think that at some point a Micro$oft marketing droid helped craft some of the words in the article - it's all just too pat.
To me the main lesson of stories like these is that if you're paying serious money for a serious business app you buy a turnkey solution with clear deliverables, a test plan, and a stability period. Then, when it doesn't work it's the vendor's problem to fix. It sounds like in this case the customer thought they could save money and do it themselves, and ended up finding out the hard way that they couldn't.
OK A. This could have been caused by some hardware malfuntion. Bad ram/controllers/processor/etc.... isnt to good for stability. B. Was he running a bleeding edge/unstable version of redhat? C. Most importantly of all did his suppose team actually know what they were doing?
According to Red Hat Australia the guy refused to run a diagnostic :
"We asked the customer to do a diagnostic test and the customer never responded, so it was impossible for us to address the issue," Mr McLaren says.
This was written by passive agressive person making all the right noises about objectiviity while doing everything possible to insure failure. His ultimate agenda was to do as much damage to Linux from his little podium and he seems to have succeeded. When his hardware problem comes back I'm sure that he won't be writing another article about how he had to actually reformat the harddrive and reinstall just to get Windows to boot.
I never understood why a company whos downtime cost 100's of thousands a hour would worry about paying $1500 for a copy of windows.
One of the biggest OS bigots I know is a Solaris bigot.
Xix.
"Everything is adjustable, provided you have the right tools"
Any admin worth his salt knows that a kernel based OS that gives you memory dumps is more then likely a hardware issue.
What I think really happened is that he doesn't know his head from his ass, and when he came across an issue he didn't have the competance to handle, he bails and blames it on the Linux to save what little bit of integrity as an admin that he has.
Go back to Windows you pussy.
I read most of the article and what struck me is that many posters missed the point.
The Red Hat rep who was interviewed said that they requested a hardware diagnostic but never received one, which leads me to believe that it is a hardware issue, not a software issue. If that is the case, then certainly Windows would work, since it is more forgiving in certain situations where there is buggy hardware.
There are enough work-arounds for that in the MS kernel that hardware problems can go unnoticed for years without crashing until that one application or utility brings it down.
Mod up I say.
Crest Electronics fails self-administered IQ test.
Yes, my only tool is a hammer. And you're starting to look like a nail.
It seems like the admin never really wanted to be on Linux in the first place and his knowledge of Linux is highly lacking. The fact that he knew nothing more to describe his problem than "blue screen of death" shows which OS he wanted in the begining.
I've had ongoing issues like that before of random crashes spaced weeks apart (userland software problem, not OS problem). I worked with the vendor very hard and we got the issue resolved over the period of a few months. Some suggested we switch to windows. Myself and my contact at the software vendor didn't think it was a good idea. In fact, it wouldn't have been a good idea, because there was a corruption in the data itself that was crashing it. An OS switch would have been loads of time and effort, just to have the problem still be there.
The fact that he never even returned Red Hat's calls leads me to believe he really didn't want the problem fixed. He wanted to make Linux look as bad as possible to his superiors so he could switch to what he really wanted. I doubt the whole operating system crashed. A misconfigured SAP was probably crashing and he was too incompetant to be able to tell the difference.
Also, what lameass autopatches on a mission critical server? That's such an incredibly bad idea. I'm sure all Red Hat's patches are of the highest quality, but if downtime could be a problem at all, take 20 minutes out of your day to look over the patches and make sure none will conflict with your particular setup. There's no replacement for human intervention if it's that important.
Ultimately I highly doubt the problems are rooted in Red Hat or SAP. They are rooted in a stubborn admin who didn't know what he's doing on Linux and found it easier to blame everyone else.
If an officer ever threatens to taze you, say you have a pacemaker.
In one sentence the guy's saying SAP won't support Linux unless it is manually patched... then the next he's talking about using auto update service for Windows?
... ?
What the
Or as another famous Queenslander once said, "Please explain?"
yes, it still uses messages to windows, as does most of win32 things, included COM.
I think what has happened is some app had a lock on the clipboard, and it is not releasing it. A crashed app should be dealt with, but an app that hasnt crashed? reboot time, I guess. Or I write the code to get teh clipboard owner and WM_DESTROY it, but that would go too near those clipboard APIs.
Response header on their webserver reports:
Apache/2.0.48 (Fedora)
Guess we know how REALLY smart their IT 'guru' really is, when he doesn't even know the platforms they rely on. All the while, shooting off at the mouth of how Microsoft saved the day for him. (oh and by 'saving the day' I mean that his Microsoft VAR greased the hole first)
How about just also using Fedora for SAP, okay?
(or if compatibility issues, then whitebox linux)
I, for one, will no longer buy their toothpaste.
... the phrase that flags this as a statement from a Microsoft shill. They must get a template from M$.
Exactly.
This excuse sounds like the long-time tech support "solution" used when the support person has no clue of what the problem really is.
Tech support: "Reboot your computer, if it still isn't working; give us a call back"
If you are a vendor; it is YOUR job to make your client's system work. If you really need the full diagnostics and the customer doesn't have the time/manpower/experience; then have one of your own people do the job.
that SAP, as a platform, sucks.
I bought a motherboard, and it stated in the instructions that Windows XP was needed for USB 2.0 support.
So, I installed Windows XP. But, the optical drives (that were just used) never showed up.
So I installed Windows 98SE. Optical drivers showed up! Until I installed the motherboard specific drivers. Then, they disappeared. And Windows 98 SE doesn't support my USB 120GB hard drive, or flash. If I turn on "DMA", the optical drives disappear under Windows 98 SE as well.
I put in a TV tuner card. It came with drivers and a PVR application. None of which work -- best as I can figure, it didn't decompress all the files correctly. So I don't know if this works with Windows (98 -or- XP). And I can't find drivers for Windows 98 for my CIMR-100 remote.
-AND- Windows 98 SE crashes on exit. Every time.
I put Linux on. Fedora Core 2 (yes, its obsolete, but I happen to have it on DVD).
USB 2.0 - check
USB hard drive - check
USB flash - check
TV Tuner - (needed installation of XAWTV) - check
PVR - MythTV - check
CIMR-100 - check
DMA to hard drive - check
All optical drives available - check
No crash on exit - check
And this is with random "craptastic" hardware. I was attempting building a PVR with the cheapest hardware possible (motherboard, tuner, video). No thought to "compatibility" with a particular OS. Purely price/performance for the task at hand.
Linux hardware support RULES.
Ratboy
Just another "Cubible(sic) Joe" 2 17 3061
This is Some kinda joke.. I've been using Linux for over 6 years. In the Past three i have seen Significant improvements. But what i have never experienced is complete lockups or random crashes. Maybe because i use "good" hardware and don't Skimp on anything. My systems are always up and running. This has to be a marketing ploy.. Lets see how much Mr. Bullshitter saved by going with microsoft. What he's not answering the phone? Oh it's not ringing because they didn't pay the bills.
My ex-wife, shortyly before our diviorce, bought an HP laptop running Windows XP. It had an S-vodeo output which we plugged into the TV to use the laptop as a DVD player.
When the PC was plugged into the TV it blue screened all the time.
She always cursed Microsoft about it, but my take was that we didn't buy a Windows computer, we bought an HP computer. It was up to HP to make sure their hardware would work with the OS that HP chose.
In the case of TFA, IBM should have made it work. It was their hardware, and they sold it with the understanding that it would be running SAP on Linux.
That said, I would never trust Red Hat, having used Novell's other crap (God but I hate their mail client)
And the counter-story. I used to be a Microsoft admin, and now have moved to an entirely Linux shop. The reason? Even a fresh install would do crazy things like packet our network (good old M$ SQL Server). The system would crash at least once a month... and if it doesn't, it would need to be rebooted twice a month anyway for 'crucial security updates'. The system could be managed by "any grade-A idiot"- of course the difficulty being that everyone else tried to change things they shouldn't (BTW: this is the biggest reason for misconfigured Windows servers- non-admins playing make-believe).
Compare this to a rock-solid 'just works' situation with our Linux clusters. I haven't had to fix a problem in years. Things just keep working the way they should rather than some new way automatically!
Maybe the problem is that a bunch of Windows admins knowing nothing about Linux started using it without any required training or valuable time on a test-bench finding the quirks, tuning, and learning the system. This is worse than the MindCraft tests where they get M$ to configure and optimize their server and then install a bare RedHat installation and send a few million requests its way.
Please. With knowledge, training, and staff, Linux is a dream. However it does take a new learning process on the UNIX side that most admins don't have... hence failure. It should _NEVER_ take 80 hours (2w*40h) to install a program- you're obviously doing something wrong.
-M
when you see the word 'Linux', drink!
I got all the management at the company praying for a switch to linux. Our first phase roll out is dieing. The servers are all win 2k and a nightly reboot is already required to maintain some sort of performance. Of course the windows admins don't know what the problem is because they know nothing about systems or troublshooting problems. I have been watching them for weeks grabbing at straws now they are upgrading windows hoping and praying it will help. Lets see how he feels about those windows servers in say 6 mos.
Several things in that "article" jumped out at me right off. I usually read things like that two or three times to find the spin points and to think about whats being said while not being said. Given that, here are a few things that caught my attention on the first pass, and bug me even more the second time I read the article...
."
First, this: "Despite using a version of Linux certified by SAP and SAP-certified IBM servers, stability issues and the complexities of keeping Linux up to date and secure. .
I have run various flavors of Linux since I first downloaded an old version of Slackware off a local BBS way back when. Another post mentioned something that I agree with. I do soemtimes forget exactly how or why I configured something on my linux machines at home and at work because I configure them and they run. Period. I dont have to touch them again. My windows machine however, usually requires a reinstall every 18 months or so. Sometimes for an upgrade, and sometimes because it just stops working right. I dont know. Admittedly, I am not a windows person. Also, I can say, having supported end users and corporate customers on Red Hat Linux, administered RH machines, tested them professionally, and so forth, that the vast majority of stability issues I find fit into two groups. A: Issues with the OS. This almost always occurs, however, when I am testing beta releases. Almost never happens with a released version of RHEL, or even SLES. B: Hardware. Either bad hardware (usually attributed to a failing hard disk, a BIOS issue with a piece of hardware or a piece of ECC RAM that starts throwing double bit errors). Beyond that, I hardly ever see a stability issue with any linux that was not caused directly by me because of something I tweaked too much, or some piece of software that I installed that had issues that have nothing to do with the underlying OS.
The next thing that jumped out at me was this: "Crest's IT manager, Anthony Horton, oversaw the deployment of SAP on Linux in November 2004, after inheriting the decision when he took the job. Having previously run SAP on AIX - IBM's version of Unix - Horton was comfortable with deploying such a mission-critical application on Linux."
So what they are saying then, is that the IT manager DID NOT CHOOSE LINUX. He came into his job AFTER the decision was made and had been initiated. It does not take a project management genius to know that switching owners mid-stream during a major project is a Very Bad Thing[tm] in most cases. Also, since he "inheirited" the project, he had no real investment in it to begin with. His own statements later on point to him being a Windows guy. Kinda like those people who dislike minorities, but try to qualify themselves by saying "I have friends, but..." And running SAP on AIX makes it comfortable DEPLOYING SAP on Linux? Linux ain't Unix. Thats like saying I ran XX software on Windows, so I am completely comfortable running XX on MacOS because they both have point-and-click GUIs. Moron.
He then says he called contractors to install RHEL and ensure it was SAP compliant??? Umm, if he was so comfortable with it, why did they farm out the installation? Installing RHEL 3 is NOT difficult. I do it several times a day on test machines of all flavors. Its NOT DIFFICULT. If they cant even install the OS themselves, complaining later on about the "Difficulty" of keeping it up do date is superfluous. Another analogy: My car wont start. Did you put gas in the tank? Gas? You have to put gas in it?
The next complaint is that it took two full weeks to get everything set so that SAP would support the box and the software. Then only two days for Windows. Ummm.. duh. The hard work (ensuring supported hardware and components) was already done when they initially set up the RHEL box. Unless they just chucked the server out the door and configured a brand new one for the Windows install, which I really doubt, they didnt have to do even 1/4 of the work during the Win
"Our funds have never taken part in toxic or death spiral convertible financings of any sort" -BayStar's managing partne
I have installed SAPDB on Suse Linux in a couple hours and had no real problems for over a year. I wonder if the hardware or Installer (person) is the problem in this case. I don't realy use Red Hat Enterprise so I can't say one way or another about it, but I do use CENTOS which is based on Red Hat Enterprise.
SAPDB also installs fine on Mandrake and Debian.
*Shrug*
My 2 cents.
Who's flying this thing? Oh, right, that would be me --FireFly
Hey, did they have anything to do with that tooth microphone I was just reading about. *rimshot*
"With the manual process of patching, we were spending about two days a month ensuring that and testing. A lot of people call it a soft cost, because you've got IT people anyway but they shouldn't be spending all day maintaining the system," Mr Horton says. So there's no need to test the automatic updates from Windows .... just the manual ones from RH. If that's the case, they should have just put yum or rhn on cron - no testing required!
What does he expect if he refuses to run the test stuff? Did he think operating systems were a solved problem and Linux was the full solution?
Come on buddy.
He's probably got a bad sector on a disk or a memory error, or something misconfigured....
I know, from personal experience, RHEL is rock solid in general.
If he doesn't buy new hardware he'll probably see similar blue screens in Windows.
In other news, a formula 1 race car was returned to the dealer yesterday by Grandma Langsam because it was "too fast around turns".
The right tools in the wrong hands will always lead to crap like this. There is no new information in this article.
Windows does run better on flaky HW such as marginal memory or memory timing. I would have thrown a different box at the problem before making the switch.
Company dials support number.
Kid: Uh hello?
Company: Hi I need support on this FreeBSD box I installed.
Kid: Uh did you RTFM yet?
Company: I'm sorry...?
Kid: RTFM first dude. Oh and by the way this is just for ordering burned CDs of FreeBSD. The support is on the mailing list. Later. *click*
Company: Ok time to install Microsoft again. Cheaper than hiring one of these damned kids.
" I doubt they make a "SAP for Dummies" book, either. "
4 503758/qid=1128003645/sr=1-2/ref=sr_1_2/102-663534 8-1619342?v=glance&s=books
http://www.amazon.com/exec/obidos/tg/detail/-/076
I'm sure my company would be happy to sell them to you for a "reasonable" fee.
(And to save your Linux machines: Kick MVFS the fsck off the system and go with snapshot views only. That MVFS thing is the dumbest thing I've ever seen in a revision control program. Obviously written by someone who liked VMS but didn't understand it.)
You know you're in trouble with a revision control system which says you can't use regular Make, you have to use their Seekrit Make That Is Almost, But Not Quite, Compatible With GNU Make, Except That You Modified GNU Make To Have New Features Because That's What OSS Is All About So You Really Can't Use The Stupid Seekrit Make From The Rat Company.
We hates them. We hates their lying salespeople. Sure, they got fired for lying, but we were still stuck with the bill for the software.
If Microsoft would pay me enough, I'd say this for them too.
I've had linux running on about 20 machines since 1998. They have been up most of that time. The only time I have ever experienced instability in the systems is when there is a hardware problem. The truth is, if this guy had problems with instability, then he had bad hardware...
If everyone gave Windows the same treatment, dumping it the first time it crashed, nobody would have grown so attached to it... Why didn't they?
The manager claims to have run SAP on AIX for several years, and to have been comfortable with that. Given the minimal pricing difference between Linux/x86 and AIX/POWER for IBM bundles, why wouldn't they have gone with AIX in the first place?
Don't get me wrong -- I've found Linux to be very solid for a number of years. But when you're dealing with package compatabilities and interdependencies, sometimes a commercial OS is easier to track and maintain.
It's also not very clear whether the core problem is with SAP or RedHat. Given the engineering squads that are required to customize an SAP installation, I'd really hesitate to presume the problems couldn't have been faulty customization rather than either of the core products.
I do not fail; I succeed at finding out what does not work.
The article dovetails nicely with this report from everybody's favorite analyst Laura Didio:
9 22/bs_nf/38232/
http://news.yahoo.com/news?tmpl=story&u=/nf/20050
which includes such gems as:
It's also worth noting that as the Linux and open-source models become increasingly commercialized, they are spawning an entirely new industry of add-on products and services. And each of these products and services come with a separate price tag.
Open Source Risk Management, for example, sells an indemnification policy to protect corporations from third-party intellectual-property claims made against the use of Linux. That service costs six figures. Meanwhile, startup Black Duck sells an automated software-compliance management tool designed to help corporate developers avoid copyright and patent infringement. The retail price is over $20,000.
Because we all know you need copyright protection and indemnification due to all the legal risks with running Linux.
I would be interested in more details. I've used Linux since 97' and have never experienced anything like what was described in this article. I am also a longtime Windows NT Admin, MSP certified in 99 or so.
I use what works, I am not a Standard Bearer for any particular OS, my posts over the last 6+ years here on Slashdot will show that.
This just seems a little "odd" to me.
p.s. I've also admin'd Solaris, HPUX, AIX, Irix boxes as well. Linux distros admin'd Slackware (back in 97-98), RH (since 4.2), etc.
Awesome!
Perhaps that final sentence should read:
."
"The total cost of ownership is actually lower in this case than with Linux because of the hidden costs of stupidity
sounds to me like their it department applies to one of the following:
...i am linux.
1) noobs
2) noobs
3) noobs
that about wraps it up. linux = easy.
Everyone who works in Software and deals in Quality knows that SAP is a piece of JUNK!! It's crap software and it never works out of the box. It's not Linux that was the problem, it was SAP.
The same does not hold true when discussing a Gentoo install, esp if it is a stage 1 install :)
the first thing that popped into my mind is "they've got a hardware issue." I've seen Windows boxes do the exact same thing; it always has turned out to be hardware.
Other things I found suspicious in the article:
and
I've been in similar situations. Applications should *never* be able to crash the system; if they do, then either there is a kernel bug or a hardware problem. Unless SAP requires the installation of kernel-level drivers, this sounds a lot like a hardware problem to me.
Don't underestimate the power of The Source
The article smells of a Microsoftie in linux-land
First off it states that Horton 'inherited the decision'
"Crest's IT manager, Anthony Horton, oversaw the deployment of SAP on Linux in November 2004, after inheriting the decision when he took the job."
Which would imply he would've made a different decision to start with.
He goes on to imply that automated patching is a big bad no-no and they must manual patch linux:
"Software updates had to be manually installed to ensure SAP certification."
Then the article goes on to say they will use windows automated patching service "WSUS":
Crest Electronics is trialling Microsoft's Windows Server Update Service, which allows automatic patching for the operating system and other Microsoft software on servers and desktop machines across a corporate network. Its benefits are one of the key reasons why Mr Horton stands by his decision to switch from Linux to Windows."
The whole article stinks of Microsoft, and I wouldn't be surprised if Microsoft wasn't behind it in some way or other.
Oh, I'm going down that road right now with Apple ... try finding a flaw in something as important as a vendor's TCP/IP stack and/or ethernet driver and/or firewall implementation, and watch as they dragging their feet over fixing the issue. I suppose the only way to get results will be going public with flaws rather than going straight to the vendor?
Comment removed based on user account deletion
Secondly, there must have been some seriously misconfigured/buggy software to make the system crash so often; leading me to believe that Linux is not at the core of their problem.
In the 10 years I've been working with Linux, I've had no more than 5 or 6 server crashes (total). (As of writing, one of our Linux servers has a little over three years of uptime. Not bad.)
A reoccurring theme that I've noticed with companies that are balking at the prospect of Linux has been the added cost of bringing in consultants to show the IT staff how to operate and setup the new software. In these instances, I would suggest either sending the Network Engineers, Systems Administrators and Support staff to a couple of training courses or better yet, go with a plug and play solution and bring on an already experienced Linux professional. If your network runs on Linux it wouldn't make much sense to have an IT staff full of Microsoft and Novell guys. Analogously, if your car operates on gasoline, it wouldn't make much sense to fill up the gas tank with diesel.
We run IBM's portal software (Java-based) and it needs to be restarted usually on the basis of about once a week or more. I love IBM for hardware and their 'nix support, but in some of their software they need to get a bit more serious about stablity and user-friendliness.
I heard Nokia located a bug in Red Hat kernel and had to wait months for Red Hat to fix it. If they had done it themselves they would have violated their support contract....
The article is unclear. The last couple of paragraphs leave the distinct impression that the problem is that SAP don't provide patches to Linux vendors (in this case Red Hat) in time. There's no reason why Red Hat's Satellite server couldn't be rolling out the patches to affected machines as needed if they were supplied in time. Either the problem is with SAP or with Red Hat not taking patches from SAP quickly enough.
Whichever it is this article is unsatifsying because it doesn't explain the issue clearly.
The uptime says it all for me. I don't have a linux box 29 days, and I have never had one stop, period. Software runs thru hardware problems that would be totally undiagnosed on windows, windows would simply stop. On linux I can even diagnose those symptoms. If this group had these problems, they will have them with windows. as the first poster said, wait 7 months, and then see. Silly posting, and obviously a Redmond troll to waste /. bandwidth.
I wonder what they had to pay to get windows up and running, and what assistance they had from Redmond, that other windoze users would not have had.
Also, I wonder what response the got from SAP when they called with problems, vs when they called with SAP install problems on the windows problems.
All relevant and not mentioned here.
http://informationweek.com/story/showArticle.jhtml ?articleID=171201387
Why can't this [relaitvely] little company in Oz make it work? Might it have soemthing to do with not running the diagnostic requested by RedHat Aus.?
I notice that he is still using Linux for lots of other stuff where it works just fine and isn't too difficult or too expensive to use.
That is actually perfect as an analogy to the real world. Windows is the hammer used to break itself :D :b
I am the Barber of Seville.
"And who in their right minds lets any mission critical server auto-patch itself, regardless of operating system. That's just utter madness!"
l ?articleID=171201387/
I can't get back to the article at "The Age" now, I'm getting a registration page and I'm not interested in registering so that I can copy and paste the pertinent quote. YMMV. They *might* be slashdotted.
But in the article he states that he has his W2k3 server set up to autopatch now that he has taken the RedHat server down. One of his complaints was that he couldn't automate updates on the RedHat box because they needed to compare the version of the patches with the compatible versions listed on the SAP website and it was too labor intensive. In my experience Unix (yes, I know we're discussing Linux here) has always been that way and so is any mission-critical application or OS. Or hardware for that matter. Use the wrong solution, get the wrong result.
Labor intensive repetitive tasks are the perfect place to implement scripted operations. He may need to train or hire someone to write a script to do this, but before that can happen he has to understand that it's what he needs in order to save himself time, money and man-hours. It's a conceptual gap which he may not be able to bridge.
As I posted elsewhere in response to the article, TRW is running SAP-based ERP on Dell servers and it's running fine and saving them bundles.
http://informationweek.com/story/showArticle.jhtm
On their website Dell offers SuSE and RedHat. No mention in the article of which one TRW uses.
((((this saves folks the trouble of clicking Read more...)))) --- . though perhaps your 5.4-P6 system doesn't have any of those programs. And note that this only includes stuff that's installed by default -- my command didn't scan under /usr/local, for example.
(Though to be fair, some man pages mention GNU or the GPL without actually being GNU software. `fdisk' is an example of that. And somebody can be released under the GPL but not be GNU software. But most of these programs are GNU ...)
is because of the GNU userland alongside the Linux kernel
Only a portion of the userland code in a Linux distribution comes from the GNU project. In fact, several years ago I saw that somebody worked it out, and it was like 30%, with the Xfree86 project contributing even more than 30%. (Granted, this was a while ago, so the exact percentages will have changed. But even so, I suspect that GNU still isn't even a majority.)
... look even worst, because are sugarcoating a reality that is bitting most computer users out there. Praising Windows goes against popular wisdom and experience.
Look pal, we have bunches of W2K machines in our desktops, and simply do not recognize your rosy scenario of always running, never failing Windows desktops.
The situation is so bad (with properly locked down machines mind you, your Linux interns would had have no chance in hell to screw our machines) that our team of very capable Windows SAs has stated as policy to fucking shutting down our desktops daily in order to minimize problems. That is how desperate they are.
Should I mention that 3 or 4 times a year our corporate network goes down to its knees due to the latest exploits of the script kiddies out there (and before you pass summary judgmenet about why we are not patching, hold your horses, while Solaris and Linux are pretty much trouble frre in the patching arena, Windows patches can completely mess up a system, reason for which testing of patches before releasing them in production takes much longer. Cost that in your TCO Windows marketroids) unleashing the power of some of our WIndows machines as spambots?
The Linux/Solaris crowd just laugh our heart and look at our reliable machines that in most case are not shutdown more than 3 or 4 times per year when major rounds of patching take place. Wisely we run as many applications as possible there and while our colleagues are going through the last round of patching and UAT testing, we keep working.
.... you can't pass such a summary judgment.
Many systems take not weeks, but months or years to be put in place.
And it is not because the providers are incompetent, but because some IT solutions are too complex and have to be finely tailored the the client's needs.
7 months to implement a system involving SAP seems to me like par of the course, SAP is not MS Draw you know....
IANAL but write like a drunk one.
Techies talking about technical stuff in a website devoted to technical matters is navel gazing.
We should all start talking about TCO and grew up. ANd throw up.
IANAL but write like a drunk one.
Oooooh yeah. You didn't get the full OpenServer rant ;-)
... not a big fan. The fact is, however, that the core OS is still extremely stable. I find OpenServer decent enough, print services aside, so long as I never need to actually *change* anything. I'll take Linux any day, but it's definitely not up there in the stability or "just put it in and forget about it" admin requirements stakes.
It may be stable, and well documented, but that's about where the good bits end. Yes, the scoadmin tools are buggy as hell. On the other hand, they exist, and frankly that's better than any Linux I've ever used. I also dislike, however, the fact that the scoadmin tools hide what they really do from you - there's precious little information about what files a tool will affect, etc.
OpenServer's `lpd' is pure evil - I smell sulphur whenever I have to work with it. It's hard to configure, quirky, unreliable, archaic, and the admin tools for it are way more broken than for any other part of the scoadmin suite. It's currently making 800 DNS requests per second to my local DNS server. The conversation goes something like this:
SCO: what's my hostname? My IP is 10.0.0.8 .
DNS: alder.localnet.
SCO: what's my hostname? My IP is 10.0.0.8 .
DNS: alder.localnet, just like the last bajillion times.
SCO: What's my hostname? My IP is 10.0.0.8 .
DNS: *shoots self in head*
Additionally, the lpd service tends to silently quit, or to start two copies of its self then stop working. That particular bug has been confirmed by all other SCO admins I've met, too. Alas, their compiler suite seems archaic and it's a nightmare to get it to work, so I haven't been able to build lprng or god forbid CUPS as a replacement print service.
So yeah
I use Windows, Mac OSX and Windows XP in equal amount, and in my opinion you've hardly used Linux at all. Adding more and more applications has nothing like the affect it does on Windows, I don't know why this is, but it is simply the case. On Windows you need to reinstall after a while to get your performance back, on Linux you can hardly tell. I can't really comment on the Mac since it's on a separate box, so the comparison would be unfair, but Linux and Windows XP both run on the same machine.
Disclaimer: this is my experience. I've never discussed this particular issue with anyone else, so maybe I'm the only one in the world who gets the described behavior.
This is has been said before but it takes a LOT of time and money to setup SAP on any system. I worked at a very large america aerospace company and have heard from a number of people that to move the old accounting software over to SAP will cost somewhere around $ 100M. Now don't tell me this will not happen without problems. Sounds like this guy is looking for a quick fix.
there are packages available.
i have personally compiled tomcat from source for use with apache on linux and i don't remember having to write any startup scripts...
sum.zero