The Fedora Project, sponsored by Red Hat, releases version 26 of their operating system. In the past I was an active desktop user. The OS offers latest-and-greatest with 6-months release cycle and 18 months life-cycle. More important: Fedora releases are the base of Red Hat Enterprise LInux releases, which are the base of CentOS releases. And that’s why I mention this Fedora release in particular. Since Red Hat is busy releasing v7.4 of her Enterprise Linux, I guess engineers are also looking at the horizon to work on RHEL 8.0. I’m not the only one with this view. Fedora 26 could be the foundation of this OS. So let’s compare RHEL 7 with Fedora 26.

So what’s new in Fedora 26 compared to RHEL7:

  • Yum is gone. Welcome DNF. In Fedora 26, DNF is rebased to v2.x;
  • Anaconda had a new partitioner tool, including support for this provisioned LVM;
  • Python is v3.6 by default. So all scripts are rebased from v2.x to v3.x;
  • The old GCC6 compiler is gone. Welcome GCC7;
  • Better (local) caching of users and groups using SSSD. A must for enterprises;
  • OpenSSL 1.1.0. Which is required to support HTTP/2 (ALPN support);

So yes. I’m definitely going to test-drive Fedora 26 and gain hands-on experience with some features like DNF and HTTP/2 which is much, much faster for SSL-secured websites, which is more common these days due to the Let’s Encrypt initiative.

If Fedora 26 will be the base for RHEL 8… I’m not sure. One of the open issues for a new Enterprise-graded Linux is Long Term Support. And the file-system is changing every RHEL-version. Was it ext3 in RHEL5, it became ext4 in RHEL6, which became xfs in RHEL7. Many out there hope that ZFS will come to Linux but licensing does not allow binaries to be distributed. Btrfs is a good alternative candidate, in particular the checksumming function which is missing in XFS but Red Hat is deprecating btrfs in RHEL7.4. A good candidate would be bcachefs. Tools for this storage type will be made available in Fedora 28, released mid 2018.