Oracle Tries To Capture CentOS Users

Home » CentOS » Oracle Tries To Capture CentOS Users
CentOS 62 Comments

Hello CentOS Guys :-)

Oracle has created a new Website to capture CentOS Users :o

http://linux.oracle.com/switch/CentOS/

62 thoughts on - Oracle Tries To Capture CentOS Users

  • Chris wrote:

    Yup. No, don’t switch from CentOS to RHEL, switch to *us*, and pay even more for very, very, very lousy support.*

    And I *love* the chart… never mind that RH changed the build structure to give Oracle pain, and it hurt us, Never mind that the difference between Oracle and CentOS, once that was fixed, dropped like a rock, due to our team (that is, Johnny, and Karanbir, and all the rest), to where it’s even….

    No, you should pay us, says Larry, I need to keep my yacht and fighter jet taken care of….

    mark

    * Long rant of personally-experienced month of Sun/Oracle “support”
    available upon request.

  • I found the “update delay” graph laughable. They’re comparing themselves to a bunch of volunteers and then say “we beat people work for free and for the love of it”…. Right… IMHO, I’d rather go for RHEL if I’m paying. (Unfortunately I don’t make that decision, customers do, disclaimer, $dayjob has OEL customers too).

  • I have two AMD Opteron servers, a self-built one with Asus mobo, and a Sun w1100z (ages old, single core Opteron). I plan to run Oracle’s on the latter, while keeping CentOS on the Asus one.

    Came here asking about OpenJDK 7 support in CentOS 6.3, as I plan to run a Java based app server on it. Since I haven’t installed Oracle’s yet on the Sun box, no need still to ask any questions (I don’t even know if there are mailing lists for Oracle’s or web based forums).

    I want to run both, and don’t see the point of having any grudges against any particular distro. But hey, that’s me.

    FC

  • Yes, pick the one bad time for CentOS and release 6 and show that in a graph. I find it interesting the CentOS is showing as faster than Oracle now.

    Meanwhile, if this is linux.oracle.com, shouldn’t Oracle database be included in the ‘free version’? If you want something laughable. Anyway, Oracle came with some old distro or book I picked up. It was free with some strings as I remember, but that is likely a decade ago now. Still, shouldn’t Oracle linux include Oracle? The word Oracle being synonymous with one thing. I might have to do an install if it did and if it was free for any use. ;)

  • Fernando Cassia wrote:

    Oh, as I read this, I was thinking, and realized I *do* have something against OUL specifically: of commercial providers of Linux, Novell’s SuSE
    is their own distro. RHEL is their own distro. Oracle takes RHEL and customizes it, rather than building it from the F/OSS components, the way SuSE and RH do, and sells it as their distro. *That’s* what frosts me.

    mark

  • Craig White wrote:

    Don’t worry, just as soon as Oracle drives RH under, or buys them, they’ll crank up the prices to higher than RH now – I mean, Oracle’s a *lot*
    bigger than RH, and can afford the loss leader for a while.

    mark “and the quality of their service will drop, if that’s
    possible”

  • Yes, but the real sucky thing here was when Red Hat cut off the community that contributed the fixes to their buggy initial versions from getting free binaries of the results that became RHEL. If Oracle
    – or anyone else – can steal their support customers by providing free access to usable binaries, well, they brought it on themselves. And they still could rectify that mistake if they wanted…

  • You misspelled, once again, “leech”. It’s a simple word, you should really use it more often considering how applicable it is in this thread.

    So for all your ramblings of commercialism these past few posts you would rather use something free instead of paying? Really?

    John

    Sued for what? I can see it now:
    If you win, you get a hug from Stallman. If you lose you get 2 hugs.

    — geekoid (135745) , Slashdot, in reference to
    claims that Emacs violated the GPL.

  • Yeah…. can we please just forget this and avoid all the litigation?

    John

    Of all the preposterous assumptions of humanity over humanity, nothing exceeds most of the criticisms made on the habits of the poor by the well-housed, well-warmed, and well-fed.

    — Herman Melville (1819-1891), novelist and poet

  • Oh please. His mastery of the language is better than many native speakers.

    If I wanted to ridicule I would; this, sir, was far from it. I’m just pointing out facts from this and previous threads.

    John

    Michelle Backman – For those that think that Sarah Palin is too intellectual.
    — Bill Maher

  • Just to clarify so there are no mistaken impressions.

    Oracle are, in the form of Larry Ellison, a bunch of leeches that are profiting off the blood and sweat of others. They are doing everything they can to, quite literally, steal the customer base of Redhat. And while some among us may think this is fine, just a form of capitalism, it isn’t: the company they are trying to steal customers from is THE
    SAME COMPANY THEY ARE GETTING THE SOURCES FROM FOR THEIR OWN DISTRIBUTION.

    Sorry to shout, but it seems that a few of you want to brush this under the rug. If you want to spin it as just business as usual you are free to do so; but you’re deluding yourself and basically full of it. And please stop referencing Novell, SuSE or anyone else. This isn’t about them. This is solely about what a reprehensible company Ellison put together and their shady and downright shitty business practices.

    John

    Failure is the condiment that gives success its flavor.

    — Truman Capote (1924-1984), American writer, Portraits and Observations,
    The Essays of Truman Capote, “Self Portrait” (1972)

  • This will be my only comment to this thread.

    The issue that Oracle highlighted, that existed in the first 3 quarters of 2011, has been fixed. CentOS now has 2 full time developers and I
    have created a graph that shows the same information, in 2012 (EL6
    Kernel release times) that Oracle depicted:

    http://bit.ly/NEdAB8

    As you can see, in 2012 CentOS has delivered the Kernels 48% faster than Oracle (25 days compared to 37 days).

    CentOS also delivered the bugfix kernel for EL6 released yesterday 17
    hours faster than Oracle.

    The speedy updates are not just kernels … if you check turn around time for all packages in 2012, CentOS has been much faster than Oracle. Expect this to continue.

    Also, expect the CR repo to be used for all future point releases … not just if we encounter problems.

    While I will not be petty enough to post a Migration script from Oracle Linux to CentOS (I can’t believe they posted that … REALLY?), I will likely help anyone who asks me for help with that specific migration.

    Thanks, Johnny Hughes

  • You’ll be pleased to know (as Johnny pointed out to me the other day)
    openJDK 1.7 is fully supported by TUV in 6.3 … so it’s a fully supported part of CentOS and not just a tech preview ;)

    Johnny I know you said you were only going to do that one post but I
    hope you’re still reading the thread – thanks for the graph… it’s most enlightening.

  • That change affects Oracle because they update and modify the kernel. Since it’s not pristine + patches like other SRPMs, Oracle has a moving target and may need to manually fix each of the patches that they apply.

    CentOS simply rebuilds the package, and I’m reasonably certain that I
    recall statements from the CentOS maintainers that Red Hat’s changes don’t affect their process. I don’t think your argument holds water.

  • As I understand, that’s not exactly true. They have two kernels, UEK
    which is not even the same version as the upstream and the one that’s identical to the upstream. They don’t have to follow every single upstream patch and apply to theirs. I have the impression they have a completely separate bunch working on their own kernel & modifications. I might be mistaken…

  • To be honest, it never had any traction. At least not down here. And it goes back to the Sun days, pre-Oracle. Back in 2006 and 2007 Sun spent a lot in marketing and was down here at Linux events with a Solaris booth or offering CDs etc… nobody paid much attention to them…

    I think the market for Solaris is the big corporations that use it to run on high-end Sun/Oracle hardware (mostly Sparc), and those prefer to pay to have a commercial grade solution with support. My guess is that ORCL thought that they’d never get any ROI over ‘free’ Solaris, nor any sizeable market share gain wrt Linux…

    While some features like zones and zfs are interesting, Linux has also improved as of late to fill some of those gaps…

    FC

  • Well, if you see it the other way around, Oracle’ s -and SUSE’ s and any other competition that might come along- help puts RHAT pricing in check…

    In the end what matters is if Oracle’s kernel programmers are any better at fixing bugs than RHAT’s and if their support responds on time or not.

    For us end users with no support contract, the more heads and firms that are involved in kernel dev and linux support, the better, more chance of getting higher quality software.

    I see plenty of patches/fixes to the Linux kernel by Oracle employees… http://goo.gl/MqjCl

    FC

  • I don’ t think that would happen anytime soon. AFAIK if you check distrowatch Oracle Linux ranks #50 and CentOS ranks #8.

    Also, I read somewhere that ORCL has 8,000 paid custmers to their Linux subscription…. hardly a major player still. But like I said, the more competition wrt pricing the better.

    FC

  • I said that Oracle Database does not interest me, because MySQL /
    PostgreSQL is enough for my simple needs. And I couldn’ t afford the price of Oracle’s commercial database even if I wanted to use it.

    Now, their entry-level Linux support subscription is another thing… if I start providing commercial services on top of a Linux box I’d rather pay whatever entry level fee I can to get *some* level of support (specially wrt priority security fixes). And there ORCL linux is the most affordable compared to RHEL.

    That’s what I said and there’s no contradiction.

    FC

  • But that flies in the face-value (rather than the hidden agenda) of their statements… Oracle wouldn’t *lie* to you would they?! /s

    If someone wanted the ‘value-added services of Oracle’ they wouldn’t be using CentOS but rather paying customers already…. and if they needed those services this ‘free service to the community’ Oracle is offering would not be of help… since they’d need to pay anyway and thus it would not be free…

  • If you look it from Oracle’s perspective, this _should_ be a win-win. They already need to manage a Linux distro for their vertical solutions such as Exabyte. It makes sense for them to base it off of RedHat ‘cos their database product is already available for that platform.

    So, for them, making the OS freely available is a close-to-zero marginal cost. And, hey, they might pick up a few support contracts on the way.

    A _user_, however, is in a different boat. Unless Oracle build a community support system (like these lists), users of OL are gonna be stuck. They’re gonna have to beg for support from other areas. You can be sure that some of them will try to get support from us, here.

    I just don’t see the value proposition of OL from a user perspective.

  • I’d only consider OL if I was running Oracle software on top of it.
    “One throat to choke”, as Scott McNealy used to say.


  • Hehe, never heard that one. Cool guy Scott… too bad Sun had to go up for sale. I have my own version “a single person (or firm) to yell at”

    FC

  • Fernando Cassia wrote:

    I agree. It was a real shame when Sun went on the block. But then, back in the nineties, I loved Sun 3 and Solaris. Most of this decade, though, Linux has become even friendlier and more useful to me.

    But then, I thought it was a real cryin’ shame when DEC went down, and got eaten by Compaq (gag!). I remember DEC and the seven dwarfs, I think it was.

    mark

  • Sun JDS Linux was damn good (included Java and StarOffice preloaded, and a cool Gnome theme).

    I used (purchased!) both JDS Linux 2003 and JDS Linux R2. JDS Linux R3 was in beta by the time the Solaris militia won the internal battle / turf war and JDS Linux was canceled.

    FC

  • Gé Weijers wrote:

    The first time I was doing sysadmin work was on a sparcserver, back in
    ’95-’97, and I really, really liked that box. I agree, it was a real shame.

    Btw, back when I was working on it, we ran into an error message that *no*
    one has ever heard of before: we had a ton of young consultants, coding, testing and debugging 12-16 hrs/day, and if a debug session hung, they just killed the window. (They were on PCs running…mmm, I can’t remember if we were on Win 3 or NT, but we were using Hummingbird.) Late one night, with a ton of zombies, another guy and I spent nearly an hour killing zombies (which we could actually do), because in their debugging sessions, they saw an error message that read “spider message” (and I forget the rest of it).

    Ever heard of anything like it?

    mark

  • Yeah, but fast forward another year or two and you had an extremely buggy OS (not a problem by itself, everything was back then) and it was cheaper to go buy a faster PC and run linux than it would have been to get the bugfixes for the OS. What did they expect to happen?

  • Les Mikesell wrote:
    Not in the mid/late nineties. Sun 3 was nice and solid (I used it, along with Irix mostly, through the early/mid-nineties). Solaris 6.3 == Sun 3;
    Solaris 6.4 was the next release, and was perfectly fine and solid.

    By the late nineties, RH 5.2 was nice and solid, and I was running that at home.

    mark

  • I’ve managed development servers running pretty much all of the Solaris releases from 2.6 (SunOS 5.6) to Solaris 10 (SunOS 5.10), and if there were bugs, they sure didn’t affect our operations. We were primarily using Solaris as a Oracle database and Java application host.

  • I was stuck with some version of Solaris 5, expired support and no free bug fixes.

    At some point, our mailserver broke and I copied the sendmail configs over to a linux box, restored some backups to it and never looked back. Wasn’t really happy with RH until 7.3, but that was rock solid.

  • Anything closely related to SysVr4 had to be buggy in the early days. It was a fantastic design but sort of thrown together. You probably stayed up to date with bugs fixed as they were found, but what did it cost you to go past Solaris 5.x?

    Linux wasn’t all that great at Java then – you made the right choice.

  • early Linux was absolutely wretched at storage management if you had more than a couple direct attached disks. fiberchannel multipathing?
    hah. ext2fs fsck on reboot after a crash when you have dozens of SAN
    volumes totaling a few terabytes ? meh.

  • There was no Solaris 6.3 or 6.4. “Sun 3” was a hardware platform (pre Sparc)!

    The SysV variants went
    Solaris 2, 2.1, 2.2, 2.3, 2.4, 2.5, 2.5.1, 2.6, 7, 8, 9, 10,… splat

    FWIW, the “Solaris” name started as a marketing thing as being the OS plus window manager, or “Operating Environment”. SunOS 4 finally got rebranded as Solaris 1.x; I still have “Solaris 1.1.1” which (from the box) “contains both SunOS 4.1.3_U1 and OpenWindows Version 3_U1”. Solaris 2 was the SysV variant and has SunOS 5.x OS plus so, for example, Solaris 2.4 had a 5.4 OS and kernel, plus other stuff (eg openwindows, deskset etc etc). You can still see it in “uname” output, today.

    Solaris 2.4 was mostly stable (I hit a few mbuf issues on Netra 5’s). Solaris 2.5.1 ran pretty well (I used it on Ultra 1 and Ultra 1+ machines, then Ultra 2’s). Solaris 2.6 worked well. Solaris 7 was a disaster. Solaris 8 was very stable. Solaris 9 I kinda skipped. Solaris 10 was too little, too late.

    (In my basement I have Solaris 1.1.1, 2.4, 2.5, 2.5.1, 6, 7, 8, 9, 10; all but 2.5.1 are original in-box distributions)

  • You’re the right man, then, whom I should ask about the elusive Sun Bandwidth Manager (bwmgr) that someone at Sun Micro once mailed to me on a shiny CD and that I manged to lose while moving. I wonder if someone might have a spare CD that could show up on eBay someday… ;)

    Apparently nobody at ORCL knows about the elusive stand-alone product after bandwidth management was integrated into later Solaris OS
    versions by Sun…

    FC

  • You mention ext2fs and I get cold sweath down my spine… I lost an awful lot of data due to ext2 fsck…

    FC

  • Thanks anyway for replying. I lose nothing by asking every former Sun employee I run across. :))

    I once built a small mini-ITX AMD x86 box mobo with riser card and one intel quad-port Ethernet card I picked up real cheap on eBay. At that time Solaris7 x86 was just released so I bought it (I still have it boxed), My plan was to use that system as a bandwidth manager and traffic shaper for my home LAN.

    Sun BwMGR looked like a great product (looking at the spec sheet at least ;), all I remember is the Sun guy whom mailed me the CD telling me “you´ll have to use the command line to configure it” because the
    ¨damn GUI¨ (sic) was coded for Java 1.x and Java 2 (at the time) had issues with it so allegedly Sun was in the process of ´re-doing it´. Go figure.

    In the meantime I moved and lost the install cd, so I didn´t even get a chance to try it.

    Anyway… I guess nowadays I could do the same with CentOS and some piece of FOSS…

    At the time, the only comparable product was one commercial solution for Linux and BSDs that was not only very expensive but also license-locked to the mac address of the adapter(s) used…. which was very annoying.

    It s been so long ago that I forgot the name…

    oh yes, thanks Google… ETBWMGR …. they´re still around… http://www.etinc.com/?p=69-ETBWMGR-Features

    Well, Sun BWMGR was comparable to that. And apparently they killed it… (then one wonders why Sun went under :-/, it wasn´t just Microsoft -and they worked hard for that).

    Afther Sol7 they bundled TCPIP QOS features into Solaris (I wonder if that code made it to OpenSolaris/OpenIndiana?) but I never understood if the Sol9 QOS features could do the kind of traffic shaping on its own without the cooperation of QOS aware routers on the LAN, as SunBWMGR used to do.

    Anyway, off-topic for this list… I know. Or not, if someone jumps in with comments about traffic shaping on CentOS.. :)

    FC

    During times of Universal Deceit, telling the truth becomes a revolutionary act
    – George Orwell

  • on 7/25/2012 6:44 AM Brian Mathis spake the following:
    Sometimes you have to share the memories before they are gone for ever…

  • I suggest death by stoning to anyone who dares to engage into light chat about OS history while conversation drifts from the original topic…

    *sarcasm*
    FC

  • Once again, the problem seems to be hatred of Ellison more than anything else.

    Red Hat and CentOS based on it are #1, it´s only natural that smaller
    players want a piece of the pie.

    This is different from Novell trying to lure RHEL users exactly how?
    http://news.techworld.com/operating-systems/106971/novell-looks-to-lure-red-hat-users/

    What Oracle is saying is “if you want free, you can use ours instead
    of CentOS and get a more recent kernel”

    centos 6.3 -2.6.32-279.1.1.el6
    uek
    http://public-yum.oracle.com/repo/OracleLinux/OL6/latest/i386/kernel-uek-2.6.32-300.29.2.el6uek.i686.rpm

    http://en.wikipedia.org/wiki/Oracle_Linux#RHEL_compatibility

    Oracle distributes Oracle Linux with two kernels:

    Red Hat Compatible Kernel – identical to the kernel shipped in Red
    Hat Enterprise Linux (RHEL).
    Unbreakable Enterprise Kernel – based on a later Linux 2.6-series
    kernel, with Oracle’s own enhancements for OLTP, InfiniBand, and SSD
    disk access, NUMA-optimizations, Reliable Datagram Sockets (RDS),
    async I/O, OCFS2, and networking.[8][9]

    https://oss.oracle.com/ol6/docs/RELEASE-NOTES-UEK2-en.html


    New Btrfs features/functionality

    An updated version of btrfsfsck, a tool to check and repair a
    Btrfs file system, is now included in the btrfs-progs package. This
    new btrfsck now supports a –repair option that allows fixing errors
    in the extent allocation tree and block group accounting. btrfsck also
    provides the option –init-csum-tree which replaces the check-sum root
    with an empty one. This will clear out the CRCs but allows the
    file-system to be mounted with the mount option nodatasum.

    Automatic defragmentation: Brtfs now provides an online
    defragmentation facility that reorganizes data into contiguous chunks
    wherever possible to create larger sections of available disk space
    and improve read and write performance.

    (snip)

    Xen domU improvements

    Several bug fixes and improvements have been incorporated to make the
    Unbreakable Enterprise Kernel scale and cooperate better as a guest
    (domU) in Oracle VM and Xen.

    Xen block backend from Linux 3.3 kernel. This provides the fully
    featured Xen blkback along with extra features, such as passing
    through a flush (a lighter version of barrier), discard (also known as
    TRIM or SCSI UNMAP) and various bug-fixes and enhancments.
    Xen PCI backend from Linux 3.3 kernel, this includes the option to
    specify how the PCI structure shows up in the PV guest – either as in
    host or virtualized; Fixes to make it work with SR-IOV VF cards; and
    numerous mutex fixes.
    Memory self-ballooning – allows the guest to automatically balloon
    depending on the workload.
    Transcendent memory support for HVM and PV guests
    Tracing API support for Xen MMU operations.
    Syncing the wall-clock time from the initial domain
    Numerous code cleanups and bug fixes (e.g. in the following areas:
    memory balloning, blkfront, P2M, E820, IRQ, MMU, Gntalloc driver)

    Other improvements

    dm-nfs: device-mapper target that allows you to treat an NFS file
    as a block device. It provides loopback-style emulation of a block
    device using a regular file as backing storage. The backing file
    resides on a remote system and is accessed via the NFS protocol.

  • Have you seen any of my other posts, where I describe Sun/Oracle’s utterly abysmal and unacceptable support? Perhaps you might, and then you might understand why I, personally, dislike Oracle. They’ve also always been the high-priced spread. I’ve never been involved with the contracts, so I simply don’t know: perhaps someone here can tell us the difference between the costs of the 3? 4? SLA contracts from RH and Oracle? I’ll wager Oracle’s is comparable or higher.

    In the US: at least five years ago, SuSE was the bigger share in Europe.

    Just one last question: if you think this highly of OUL, why are you here? Why haven’t you gone to their lists?

    mark “not imputing someone being a marketing troll”

  • No, as IBM doesn´t make DB2 GPL either. It´s the big bucks from the
    fortune 500 proprietary products which pays -in part- for all the FOSS
    goodness like OpenJDK, Netbeans, Virtualbox, Glassfish, MySQL, Btrfs,
    InnoDB, BerkeleyDB…

    If you mean, ´free as in free beer” … I guess it´s a matter of
    keeping things simple… as they try to mirror the RHEL package set…

    Plus, I wouldn´t pay for Oracle DB, with all the free options
    (including MySQL community edition) available on repos…

    Just my $0.02…
    FC

  • Um, there *is* a free version of Oracle. There are coded-in limits to size
    – I know, because in ’09, I was fighting to install and configure
    spacewalk (*bleah!*), which *required* Oracle at the time, and the free
    version of Oracle would only allow a max mem of 1G, and just to make
    spacewalk work, I had to up the default to something like 997M.

    mark

  • I think John was saying that Oracle isn’t trying to “mirror” the RHEL
    package set, but “leech” off the RHEL package set.

LEAVE A COMMENT