Slashdot Mirror
Red Hat CEO: Bring On the Clones
An anonymous reader writes "Best Buy and Barnes and Noble have a problem with showrooming — shoppers checking out the merchandise in their stores and then proceeding to order the goods at a discounted prices online. And Red Hat might have a similar problem with people (not just college kids and software professionals boning up on their skills at home, either) using the free-as-in-beer CentOS rather than licensing Red Hat Enterprise Linux and paying support fees. But according to CEO Jim Whitehurst, Red Hat's competitive position may actually be helped by CentOS in the same way that counterfeit Windows products sold on the streets in the Far East may have helped Microsoft — by cementing their position as the technology standard, in a marketplace that also includes entrants from SuSE, Debian, Oracle, and Ubuntu, just among Linux-based entrants. Who does Whitehurst consider to be Red Hat's most direct threat? VMWare."
Downloading CentOS isn't at all like pirating a copy of Windows--Red Hat consists almost entirely of open source code. People pay for Red Hat for the support. I've actually worked on a cluster where we paid for one copy of Red Hat for the head node, then loaded 15 copies of CentOS onto the remaining nodes. Nothing wrong with that at all.
Admins never needed vendor support, managers do. That means that CentOS trains the admins on Red Hat and then managers pay for the supported thingie.
Because it is not geeks that pay for Red Hat but their bosses.
The only way I can get by using my IT mandated RedHat box is by installing CentOS packages on it. RedHat simply doesn't keep the packages I need up to date. If CentOS didn't exist, I wouldn't use RedHat at all, which would entail a huge fight with IT. Thanks CentOS!
Give me Classic Slashdot or give me death!
When you have servers labeled production, that generate revenue and downtime means lost revenue, then you pay for support since its cheaper than losing revenue and customers
CentOS, which is COMPLETELY legal and above board, has absolutely nothing whatsoever in common with counterfeit Windows products.
CentOS:
1) Violates NO copyrights
2) Is not passing itself off as something else
3) Has never been treated by Redhat as anything but completely welcome.
4) Is produced by completely building from (libre!) source, not disk copying the install media.
5) Is careful to remove Redhat branding where trademarks are involved.
Jim Whitehurst never uttered the silly parallel as far as I can see, nor implied it. He just made the obvious point that CentOS does not hurt Redhat but may well help it.
It's hard to go wrong with RedHat. From a management perspective, it's a lower risk to just shell out for RHEL vice even thinking about something like CentOS. It's the "no one ever got fired for..." thing at work.
And as far as a company to give money too, you could do a lot worse than RedHat. They contribute a lot of stuff we don't think about.
Also support is one of those things that's undervalued by the technically minded. Yes, there is a great community around linux, and yes, a technical guy can probably find enough resources for free to solve just about any problem.. but there is something to be said about having a phone number you can call and someone is literally being paid to give you an answer.
This.
I have never used their support, but we pay for it on any server running commercial software since their license always requires it.
The virtualization itself is awesome - to my knowledge nothing proprietary beats the stability and performance of KVM. The weakness is the fancy tools around managing your virtual infrastructure. But Red Hat, OpenSUSE, and others are working very hard to make the tools built around KVM, VirtualBox, Xen, etc... better so they can compete with the best VMWare has to offer head-on.
Also, if people (not managers, but hobbyists, students, and bored IT folks on their own time) learn RHEL-like distros, then that means there are more people who are familiar with the environment. That in turn means more software targeting that environment, a bigger talent pool for companies to hire from, and greater mindshare.
Better for RedHat to 50% of enterprise Linux and 40% of those users paying than 100% of the users of a distro with only 10% of the enterprise Linux market. More marketshare is pretty much always good.
One can easily imagine a scenario where some startup hires a bunch of guys who "know RedHat" and set up servers using Cent. As they grow and start needing additional support and enterprise-targeted features, though, who are they going to turn to? Switching to RHEL is going to be less disruptive than pretty much any other option at that point, right?
There's no place I could be, since I've found Serenity...
If I had dozens or hundreds of servers and every minute of downtime could cost the company thousands or millions of dollars, I would want support professionals I trusted available at a moment's notice to fix things when they break.
My first inclination would just be to hire competent people on my own - if I'm the one paying their paycheck directly, and I treat them with respect, I would hope that a sense of loyalty and a desire to keep from collapsing the company that issues their paycheck.
But if I truly believed Red Hat, Microsoft, Canonical, HP, Dell, Oracle, or anybody else had genuinely first class support staff, I would consider having a smaller number of my own staff and relying upon the vendor as needed. My own inclination is to support Red Hat, because just about everything they do is open source. Canonical would come second, with the rest following behind and Oracle last. But that's my personal favoritism towards open source, if I was running a business it would just be a cost-benefit analysis including heavy research into the support experiences other companies have had with each vendor.
I guess there is not enough Linux geeks for every company, I guess even experts do not know everything, and as long as you can do everything, we just ask you to do more for the same price.
something comes up, you don't know how to solve it off the top of your head. quick research yields nothing. your company is losing revenue. you don't have time to post a question on a forum and wait a day or so for a solution. for that system you pay the 4 hour or less support costs so that if you need it, you call the vendor and get someone on the phone NOW.
where i work we pay Cisco and other vendors for support for this reason and the fact that with a lot of vendors you need to pay to get patches and updates
Why a company outsource security (watchmen) to a 3rd party, or office cleaning services?. Because they don't want the overhead of having to schedule people times, vacations, salary payments, hardware they need, training. Instead of tha,t they contract some service that do that for them. You take care of the people you need for your core service or product, let the rest to others
I can't say I enjoy working with RHEL (or its derivatives); I'm known for making a sour face whenever RHEL or CentOS are even mentioned. I can see why it's so popular (extensively validated rock-stable code), but these very same attributes make it very poorly suited for our needs (scientific computing - often using bleeding-edge software features and needing to squeeze the last bit of power out of bleeding-edge hardware).
But ask me about the company Red Hat? I'm a big fan of them. They have a relatively pure Open Source business model, and are showing the world that good money can be made out of it too. Not to mention their attitude. "Wanna clone our operating system? Be our guest, you'll only make us stronger."
On a more serious note, they're probably right about CentOS cementing their position. See also this very insightful post.
There are audit and compliance issues, that will prevent some workloads from EVER going into production, without support for accredited or validated configurations.
Just PCI-DSS is tough enough - if you need to walk a QSA through your homebrew hosts.
"Flyin' in just a sweet place,
Never been known to fail..."
It doesn't prevent down time.. It prevents unemployment time.
Situation: Server down due to OS security hole - revenue being lost .... Cut to executive management office
CTO: What's up with the server!? We can't sell our widgets and the CFO is saying we are in danger of not making our numbers now. We got to fix this NOW or we are all toast!
Middle Manager: Well, sir, we've contacted our OS vendor who is looking into the problem and as soon as they have a fix, we will get it installed.
OR would you rather say...
MM: Well sir, we've been on Google all day looking for the kernel developer who inserted the security bug to no avail.
I'm guessing the first one is better.. Lets you blame somebody else and keep your job. Actual down time is not prevented, in fact it's likely longer using Red Hat over competent local admins.
"File to fit, pound to insert, paint to match" - Aircraft Maintenance 101
I think he meant that anyone can get a copy of CentOS and train themselves to acquire the necessary skills, so when they need a paid-for, licensed and supported linux, they go with RHEL.
Its a bit like how Microsoft sells technet subscriptions for next to nothing, so people can play with all the toys like active directory and exchange and learn how they work with some hands-on experience. Oh wait... like how Microsoft *used* to do that, dumbasses.
I worked at a company that had that level of support from Microsoft. It cost millions. When things crashed it still took days to weeks to fix the root cause, by that time it had been worked around somehow. You can spend a fortune on support to buy nothing but a good feeling and when things fall apart and you need a fix fast you are often on your own.
where i work we pay Cisco and other vendors for support for this reason and the fact that with a lot of vendors you need to pay to get patches and updates
Paying the vendors for that reason is all well and good, but when did you actually get a solution from a major vendor for a critical problem within a day or so?
I have been in the IT business for a while, and I have never seen it happen. Yes, failed hardware can be replaced within 4 hours (although even that can be problematic enough to achieve in practice) but anything that is not just a case of escaping blue smoke?
The rolling support system where they shift the case around time zones to keep working on it 24 hours a day gets tedious as well, when you have to explain everything all over again every 7-8 hours.
Finally! A year of moderation! Ready for 2019?
Why do you assume Microsoft represents the industry?
From my understanding Redhat Support buys you direct access to not only kernel programmers but the distribution people. I've heard of situations where high dollar customers got Redhat to troubleshoot a problem and provide them a custom kernel to fix the problem and then rolled the changes into the main kernel.
Microsoft's business is selling licenses. RedHats business is selling support.
That's silly, all open source visualization is shit. no need to be threatened when you already suck.
No it isn't. Xen and KVM are both at least as capable as anything vmware has.
Amazon web services is based on xen.
Rackspace cloud uses xen too.
Linode uses xen.
Digital Ocean uses kvm.
Any of them could have used vmware if they thought it was better.
"When you have servers labeled production, that generate revenue and downtime means lost revenue, then you pay for support since its cheaper than losing revenue and customers"
MWAHAHAHAHAHAHA!
Ainnns...
Are you kidding, ain't you?
When downtime means revenue and you really want to do the proper thing you architect your systems so there's no downtime and you don't hire bottom-of-the-barrel technical staff for peanuts. No, sir, vendor support is not to avoid downtime but for the manager's CYA policy.
You can run windows inside kvm just fine
If an experiment works, something has gone wrong.
I saw a trouble ticket last year that had this line in it (I think it was after Sandy):
[date/time] Site router still down. Technician not en route, no ETA. Escalated to GiantTelCo 8th level support.
I don't care how stupid a manager is, at some point even the dumbest is going to recognize "8th level support" is bullshit. Then they finally might start asking "what are we paying for, exactly?"
I don't see a problem with the above. Obviously, the operations managers over at GientTelCo were fervently praying to whatever supreme being they believed in--because that was the only way the situation was going to get resolved in a timely manner. Given that level 1 is front line, level 2 is escalated support, level 3 is your subject matter expert, and level 4 is usually the engineer who designed the thing to begin with, level 8 is obviously almighty god. Had the guy working the ticket provided more complete information, you would not now be thinking that he was a dumb ass. :)
What part of "shall not be infringed" is so hard to understand?
"admins do not need any training"
They do. What they don't need is "payed vendor certifications" except for their manager.
"do not need to have certified hardware cause they can perfectly guess what is working just by looking on specifications"
Of course they do. Why they shouldn't? But even then, right now we are entering the 10th damn month! with heavy performance problems on a solution deployed by a big name vendor on a fully certified stack, or just few weeks ago we needed to update the full firmware on some chassis and blades because they weren't even able to power on. There your "certified hardware" goes.
Admins need working hardware, not certified hardware; for the most part a sysadmin worth his salt can figure his way out of the specs and if not, it's just a matter of testing it; if it doesn't work, you return it.
"because all admins are experts in every possible domain."
Of course not. But a competent admin, and moreso a competent admin team are experts in those domains relevant for their business and, much more important, provided proper means, are experts in becoming experts on them, usually at a price tag and with better results than a vendor "gold" support contract.
Caveat: we are talking open source here, since this news is about Red Hat and open source variants taking or not a portion on Red Hat business. Of course, closed source is a different issue and decision takers should pay attention on those differences and the respective value they bring to the table long term.
Provided that RedHat will actually support it. We have sent people to the training, set up equipment to the standard *their certified instructors taught* us, and then had support hang up on us because our disk setup wasn't according to RHEL support's standard- which apparently has nothing to do with what you learn in their certified classes. WTF?
Cisco doesn't do this. Even EMC doesn't do this. But RedHat did do this, several times. RedHat support also did a lot of the whole "it must be the hardware vendor" routine. That is why we moved off of RHEL to CentOS- better patch support, more packages, and no bill for support that we haven't been able to use. I haven't run into an issue on CentOS that I couldn't solve or work around with access to their bugtracker. Same can't be said for my experience with RHEL.
If I were to pay for RHEL or CentOS support, I would sooner pay SuSE https://www.suse.com/products/expandedsupport/frequently-asked-questions/#faq21
KVM and VirtualBox on Linux both run Windows VMs flawlessly. KVM was originally created by engineers at Qumranet for the specific purpose of running Windows guests in virtualization on Linux hosts. https://en.wikipedia.org/wiki/Qumranet
I always thought almighty god was zeroth level support. As in "You turn this sucker on and I'll pray to god that it works this time!"
From there, it's not good. You pick up the phone and descend to to the first circle of limbo, reserved for call center operators doomed to read from scripts. Next is the second level, where the phones are answered by system support groups, pummeled eternally by threats full of hot air to call their managers. The third level is a noisy, cold, icy machine room where system administrators are berated every time they take a call. The fourth level is where engineers are forced to joust with managers to keep their jobs while their phones ring endlessly. The fifth circle is where the VPs are tormented by joyless CEOs, members of the board, and majority stockholders in status meetings. The sixth circle is reserved for the salesmen who lied about their company's products, where they are surrounded by stacks of flaming four color glossy brochures touting features their systems never supported. The seventh circle is where the CFO sits in a rain of fiery charts of accounts and charges of embezzlement. The eighth circle of callbacks is where the company lawyer sits in a tarpit, where judges and prosecuters jab him with pitchforks full of product liability lawsuits. And the ninth circle is where the CEO is strangled and choked by a rack full of ethernet cables and fiber optic pipes, never understanding why none of them can ever hook up his iPhone to his PC no matter how hard he tries, and is bludgeoned by passing stockholders throwing dead batteries at him.
At least that's how we do support.
John
Entirely your fault, you should always disconnect cables before handling them; otherwise there is a risk of shock...
Be very, very careful what you put into that head, because you will never, ever get it out. - Cardinal Wolsey
I've been a paying Red Hat Support customer since before the Halloween Documents were published.
Access to their high-end people is generally being monopolized by the biggest customers at any given moment. If you have a problem, and IBM has a problem, and Dell has a problem... you're at the end of the line for access to Alan Cox. You'll probably have to settle for Nalin Dahyabhai, who has even more rudimentary social skills than Alan does.
Posting AC as I cannot sign in at my new employer due to firewall/browser configurations ...
I used to be the tech for MS who was sent to customer locations - and I can tell you there are varying levels of problems that customers have. I don't know your specific example, but I never had a problem fixing a problem quickly when it was something that MS wrote. The issue was actually re-producing the results so that we can figure out what the root-cause was. 95% of the time the customer was slowing us down because they wanted to be "safe". When you are talking some 16-bit accounting software that has been jerry rigged into a windows 2003 server and it keeps throwing kernel errors and blue screening, yeah those take time to fix. A majority of my calls were from customers who had some app and it would not work right, and caused issues. Getting those app support people to 1) admit they were at fault was the first problem and 2) getting them to fix the app was always the hard part.
To each their own, I always tried to provide the best service I could, and when I am there at 3 am fixing stuff while the customer has gone, they don't see all the stuff that is going on as well, they just want it fixed now.
I've heard of situations where high dollar customers got Redhat to troubleshoot a problem and provide them a custom kernel to fix the problem and then rolled the changes into the main kernel.
Doesn't have to be high dollar for the general response, actually, just the time. I had a not terribly critical RHEL 5 box with bare minimum support level (and we had maybe two RHEL licences in the office, both low level). Filled out the web ticket, got a call back in half a business day or so (i.e. within the SLA for my support level), sent the guy what he wanted, he got back to me in a few hours that it was a kernel bug and he offered to send me a patched kernel RPM or it'd be in the repos in a few days. Uptime didn't matter on that system at that time since the project that used it was over and no one else had started using it yet, so I told him I was ok waiting and I pulled an updated kernel to fix my problem from yum next time I checked a week later.
Also, on the previous topic of why any good admin would bother with RHEL, if you are in a situation where you need a C&A (such as a company that deals with US Federal gov data, not just DoD, as the requirement comes from OPM), I've found that some DAA's are much happier to approve software where there's a company behind it that could be sued and not just a random group of people who would be a pain to track down and are likely in multiple countries. In those sorts of environments you're wasting a lot more dollars in man hours dealing with C&A garbage to justify CentOS or Scientific Linux than the cost of RHEL support (not to mention the irritation). I've also had DAA's that are fine with FOSS (and I didn't bother with RHEL for those, though I'd consider RHEL for production servers in those environments).
My experience has been that RH support primarily gets one offshored robots. Me: Foo is broken Them: We cannot be fixing that because that would be FORKING TEH_DISTRIBUTION More than once the offshored frontline droids have given me answers that were demonstrably wrong and even dangerous. A few times with repeated insistence I've been passed along to an engineer on-continent who was able to at least understand the problem. We're looking at CentOS to avoid the licensing altogether for some systems, but I wonder about the CentOS lag: there have been a couple of times when something broke, notably the kernel, and RedHat got a fix out reasonably quickly. With CentOS, would I have to wait months for such fixes to filter down, so that I could make my systems bootable again? This would be slightly less of a concern if there were any way to revert updates. I really miss Solaris' Live Upgrade.