A big step forward in business Linux: Red Hat Enterprise Linux 7 arrives

A big step forward in business Linux: Red Hat Enterprise Linux 7 arrives

Summary: Enterprise Linux users, awake! Red Hat has finally released Red Hat Enterprise Linux 7, and it looks like it's going to run on everything, from the server in the back-room, to datacenters and the cloud.

Welcome to the newest Red Hat Enterprise Linux: RHEL 7.

Red Hat took longer than expected, but corporate Linux server, datacenter, and cloud users will be very interested to know that after over half a year in the beta oven Red Hat Enterprise Linux 7 (RHEL 7) is now fully baked and is finally in the wild.

However, the Raleigh, NC-based Linux giant wants you to see RHEL 7 as being more than just the next step forward in what's arguably the most popular business Linux server. Red Hat says that RHEL 7 "lays the foundation for the open hybrid cloud and serves enterprise workloads across converged infrastructure."

What Red Hat means by this is that, as Red Hat's president of products and technologies Paul Cormier said in a recent webcast, RHEL 7 works on four platforms: Bare metal servers, virtual machines (VM), OpenStack-based Infrastructure-as-a-Service (IaaS), and Platform-as-a-Service (PaaS) clouds. These, in turn, can be used together to form a robust, powerful datacenter and cloud environment for business.

To help make this happen, Red Hat emphasized that, while it's still using Kernel Virtual Machine (KVM) for datacenter and cloud virtualization, it's also adopting container technology so that users can get even more applications working on the same server hardware.

In particular, Red Hat's platform business vice president Jim Totton said that Red Hat has developed a close partnership with Docker, which just released Docker 1.0.

With Docker 1.0, RHEL 7, apps run independently from the underlying operating system and from one another. This means users can move their Dockerized apps from a container on bare-metal to a virtual machine to a cloud as needed.

Eventually, Totton added, there will be a lightweight version of RHEL 7 that will work with a Docker container. This "Project Atomic," is still in development. Users will see it first in its testing ground and community operating system, Fedora.

Don't mistake Red Hat's focus on containers and cloud functionality to mean that they're ignoring RHEL's core Linux features or its old customers. For example, RHEL 7 will be available for desktops and workstations. Totton also added that while RHEL 7 features would be coming to Red Hat's newly acquired CentOS, a RHEL clone, the final decision on when those new bits will be added to CentOS will be made by the CentOS's board.

As for the features, RHEL 7 boasts many stability and performance upgrades. Red Hat claims that, depending upon the load, RHEL 7 is 11 to 25 percent faster than the previous iteration of the software, RHEL 6.

Red Hat claims that RHEL 7 is significantly faster than its predecessor.

In addition, advanced programming tools are now available via the already released Red Hat Software Collections 1.1. The other top new RHEL 7 features, as I see them, are:

  • XFS is the default file system: This enables you to scale file-systems up to 500 terabytes.

  • Microsoft Active Directory (AD) support: With this you'll be able to have cross-realm trust Windows and RHEL domains. This is ideal if you have users working with heterogeneous operating system-based datacenters or server farms.

  • The adoption of systemd: This is the replacement for init, the old Unix way of starting processes and services on a system and initializing resources. After years of debate, it's been adopted by Red Hat, SuSE, Debian, and Ubuntu to become the new default way to start Linux systems. It also incorporates performance profiles and tuning and instrumentation for optimized performance and easy scalability.

  • The adoption of OpenLMI. This is a standard remote application programming interface (API). Red Hat has used this to provide unified management tools and a management framework to streamline administration and system configuration.

  • The adoption of Performance Co-Pilot, a set of real-time frameworks and services for recording and monitoring system performance. This lets both system administrators and other sub-systems, such as systemd, keep a close eye on what's actually happening in a RHEL 7 server instance as it happens.

Put it all together, and RHEL 7 looks very impressive indeed for corporate Linux users. I'm not the only one who thinks so.

In a statement, IDC's program vice president for servers and system software Al Gillen said: :

Red Hat has systematically grown the capabilities and value proposition of RHEL with each new release. RHEL 7 is no exception, and layers features and support for Linux Containers on top of an operating system that has seen major virtualization and cloud enhancements in the past two years. The addition of cross-realm trust with Active Directory is a pragmatic move, especially given the widespread use of Active Directory as a primary identity store.

RHEL's existing customers seem to see RHEL 7 a real step forward as well. Totton added that most of RHEL's users plan on on adopting RHEL 7 in the next six months. He added that 90 percent of the Fortune 500 are already using RHEL.

Leaving aside Red Hat's self-congratulations, from what I can see, and from what Red Hat customers have told me, I have no doubt that Totton is right. RHEL 7 is indeed a very significant Red Hat release and it's going to see wide adoption very quickly, not only in server rooms and datacenters, but on OpenStack-based clouds as well.

Related stories:

Topics: Enterprise Software, Cloud, Linux, Open Source, Virtualization

Kick off your day with ZDNet's daily email newsletter. It's the freshest tech news and opinion, served hot. Get it.


Log in or register to join the discussion
  • Every one at MS, abondon ship, abandon ship

    RHEL 7 is on the horizon, and Her Majesty's ship is out-classed, out-gunned, it just cannot possibly fight this new foe. Drop your weapons, raise the flag of surrender, give up, go home, and begin the process of rewriting the code of tens of thousands of programs that were originally written for Windows so that they work on the new REHL 7 platform.

    Time for MS to close up shop, get those thousands of pink slips in the mail, sell of the stock while it has any value, and give away all its patents.

    Will the last person leaving Redmond please turn off the lights.
    • Why bother?

      Windows, along with Windows server applications, run just fine with RHEL KVM.

      As for rewriting code for Linux, that's exactly what Microsoft has done with Microsoft Office for Android. Not to mention the Bing Search and Skype apps for Android. I wouldn't be surprised if Microsoft didn't have to hire additional programmers for its software for Android (and iOS).
      Rabid Howler Monkey
    • Partisan drivel

      Almost makes me think that:

      1. You're being sarcastic.
      2. You actually work for MS and are trying to give real Linux users a bad name.
      John L. Ries
      • The OP is a happy Linux Mint desktop user

        Rabid Howler Monkey
        • Millions have moved to Linux Mint because...

          ... it's decent, stable, secure, easy to install, easy to learn, easy to use - real user friendly OS with 5 years support time. Now when RHEL 7 is there i'm looking forward CentOS 7 coming. I wanna try what stable long term support RPM reliese can give for ordiany pc users.
          • Centos for ordinary pc users...

            That would be fedora.

            Now that centos flies the red hat flag, the days of it just being a clone of RHEL are over, however their focus will not be the desktop.

            As I understand it, Fedora will remain the end user/most up to date/testing bed distro in the group (comparable to linux mint).

            Centos will release special interest group releases (a bit like fedora spins) one will still be the rhel clone, but this will be supplied from red hat to them for modification, allowing centos debs to move from cleaning and rebuilding code, to development, focusing on virtualisation and the cloud space.

            Cent OS will remain targeted at the enterprise, not the ordinary desktop user.

            It's no different than comparing debian with mint - it's the stable grandfather, but most home users aren't going to have as good a time with it.
      • Neither.

        He/She's an actual Linux user, sadly...
  • I will not be moving my company to RHEL/CentOS 7

    Systemd is an abomination and is clearly not designed for the Enterprise. We'd rather run BSD than deal with anything using systemd, with its non-deterministic idiocy.
    • Systemd is a pain

      It is the direction the industry is going (OSX has something similar), but it's still a pain to configure and the tools for that purpose are downright primitive (at least what I've seen). Seriously, how hard can it be to get systemadm to display the available services in alphabetical order?

      init is hardly foolproof, but you only need a text editor to configure it, the configuration file and startup scripts are easy to read, and you don't usually have to radically change the versions provided with the system.
      John L. Ries
      • And the sysVinit scripts are reliable.

        If they work once, they will work every time.

        With systemd, you get it to work... fine.

        But adding a single service/update an existing service... and it may not work as the scheduling order changed, and a previously unknown dependency has arisen.
      • Another problem is the logging procedure...

        It used to be trivial, easily handled, even to forwarding logs on the net.

        Not anymore. The logs are in a binary file that can get corrupted if there is a crash or some failure of journald (the replacement) or a system crash. Remote logging is not possible - you have to run rsyslog in addition, which can get the log from the file as it occurs (unless in the aforementioned failures when you lose it all).

        On top of all that- shutdown now has a tendency to hang on improper dependency control...
    • You might want to try out Slackware

      From what i can see, Patrick Volkerding has no interest whatever in switching to systemd.
      John L. Ries
    • Actually, systemd was designed for the enterprise.

      Systemd has many more options for how, when, and what services are started on boot than init.d ever had. Like most anything new, it is only harder to use until one takes the time to learn how it works. Then the many advantages of systemd become obvious.

      It's no different than Gnome 3, whereby many people spent a whole 5 minutes using it and declared it un-usable. I didn't like either Gnome 3 or systemd when I first used them, but I wouldn''t go back now.

      Like I tell the people working for me: If you have an insurmountable case of "resistant to change", go work in the paper clip industry, where the only change in 50 years has been plastic coatings.
      • It's not progress...

        ...unless it represents a real improvement over the status quo. If SystemD was designed for the enterprise, that's fine, but that doesn't do anything for small businesses or home computer owners. And even so, it would be a lot less painful to configure if the tools were better.

        And the splintering of GNOME speaks for itself. At this point, I go back and forth between MATE and XFCE and fail to see how GNOME3 is an improvement over either. But if you like it, then more power to you.
        John L. Ries
        • I agree the tools for systemd could be much better.

          From my experience, as adoption of any given process increases, the tools to maintain it also improve. Nobody wants to spend time writing a good frontend for a process that nobody uses. Hopefully this holds true for systemd.

          As to Gnome 3, I didn't say I love it. I am still miffed that with Gnome 3 failback mode being replaced with fallback mode in 3.10 that requires 3D hardware. Now I must install Mate so I can use NX to remotely have a desktop on far away servers, for the once or twice a month (or year), that I need to use a GUI on said servers.
      • really?

        "Systemd has many more options for how, when, and what services are started on boot than init.d..."

        Actually not.

        With sysVinit you can configure even more than what systemd can.

        Why? because the administrator has total control over how the system starts and shuts down.

        With systemd you are limited to what systemd understands - which is far more limited than a system administrator.

        That is why people are still having trouble with it booting and shutting down. Things still get out of order.... sometimes.

        Until you have a deterministic boot and shutdown you can't have a reliable system.

        And accusing someone of being "resistant to change" is just being ignorant of how flexible the sysVinit is.
    • Aggreed

      Systemd is a solution looking for a problem. Apparently they found one on RHEL.
      Herby Stoukette
  • Does RH7 Still have the config_hz set at 1000

    for desktops? I once asked our provider about it and after contacting RH they told us that it had to do with tickless kernel which is discussed here:
    "The kernel in Red Hat Enterprise Linux 6 runs tickless: that is, it replaces the old periodic timer interrupts with on-demand interrupts. Therefore, idle CPUs are allowed to remain idle until a new task is queued for processing, and CPUs that have entered lower power states can remain in these states longer. "
    But if you find me a server in the enterprise which needs to go to sleep and not awaken annoyingly by having to do work, I will understand the tickless kernel.
    The reason you would want to have config_hz at 1000 is to have a responsive desktop, not a server which needs to be able to meet the demands imposed on its primary service and purpose.
    Debian by default selects 250HZ.
    Most of the improvements in RH7 are only improvements relative to RH6 and not improvements from the perspective of the Linux distributions; many other distros already have and have had for a while what is considered to be new in RH7.
    Yes they do nice stuff, I would love for the FreeIPA and NSS stuff and 389 to be in all the other distributions as fully functioning and operational but besides these, I curse at RH6 everyday especially when it comes to package management and availability of applications.
    They are a business and that's good for them, but I am limited at work to using it when, if my provider offered Debian or even Ubuntu, I would be able to do much better instead of spending time getting stuff that they didn't see as important to implement working.