Linux kernel 5.9 has arrived with plenty of performance boosting to intrigue data center admins. Jack Wallen highlights the improvements and shows how to install the mainline kernel.
It seems like yesterday we were hailing the biggest and most boring Linux kernel release ever with 5.8. But you know how fast time flies when you’re in the tech sector. Blink, and you might miss something crucial.
Lately, the same holds true for Linux kernel releases. The dev team has been crankin’ ’em out faster than we can install ’em.vAlthough quite the exaggeration, you get my point.
Said point is…there’s yet another new kernel release to celebrate. Huzzah.
And that’s about the extent of the rejoicing. Because, as in the case with 5.8, there really aren’t any game-changing features to be found with Linux kernel 5.9… at least not for the general user. If, however, you’re a hard-core IT pro, there are a few features that could have you clamoring to get the 5.9 kernel installed in your data center servers (more on this in a bit).
Let’s take a look at some of the more important features in the latest Linux kernel.
SEE: Linux file and directory management commands (TechRepublic Premium)
Linux kernel 5.9: Big performance boosts
The big news is the performance boost Linux kernel 5.9 brings to both AMD and Intel architecture. Intel-based systems will enjoy performance gains by way of FSGSBASE support, which is finally complete in this release. The FSGSBASE instruction was introduced with the Intel “Ivy Bridge” CPU, and increases performance under context switching heavy loads. This new instruction can also enable user space to write GSBASE without kernel interaction. The addition of FSGSBASE should give the Linux kernel impressive improvements on both benchmark tests and real-world usage.
If you have ARM/ARM64 devices in your data center, you’re not left out in the woods. The Linux kernel 5.9 brings along with it a performance boost for this particular architecture by way of optimized scheduler utilization information—Schedutil. This new addition is similar to that of Intel P-state push.
The ARM/ARM64 architecture is also receiving a new sysfs toggle, which controls CPU energy efficiency optimization, and numerous CPUfreq fixes.
Linux kernel 5.9: Storage and file system improvements
Considering Fedora 33 is migrating to Btrfs, the new kernel seems apropos. Linux kernel 5.9 sees Btrfs enjoying numerous performance enhancements. And with the addition of more support of NVMe 2.0 for SSD devices, file system performance on servers should be much improved. One important enhancement for Btrfs is the addition of a new rescue mount option, which groups all existing mount options together for more reliable recovery.
Btrfs isn’t alone in receiving patches, tweaks, and various bits of tuning. XFS, ext4, and F2FS have received plenty of love in the latest Linux kernel.
Linux kernel 5.9: Miscellaneous improvements and additions
There are a number of other enhancements and additions that will be viewed as important to Linux data center admins. The short list includes:
Better anonymous memory management;
a new slab memory controller, which should lead to lower overall kernel memory usage);
support for ZSTD-compressed kernel, ramdisk, and initramfs will lead to faster boot times;
USB 4 specification has received more support;
initial work on support for IBM POWER10 processor; and
bug fixes for race conditions and memory leaks across all modules.
How to install Linux kernel 5.9
Before I show you how to install the latest Linux kernel, know that you probably shouldn’t do so. Why? Because your data center-capable Linux distributions such as Ubuntu, RHEL, SLED, and CentOS haven’t added their optimizations into this release. So, although you might gain that performance boost and the additional new features, you might be missing out on features gained by sticking with your distribution’s default.
You might have a good reason to upgrade to this latest Linux kernel–or you might simply want to test it on a non-production machine. Either way, installing the latest Linux kernel isn’t all that hard (at least on certain distributions).
I’m going to walk you through the process of installing the Linux 5.9 (mainline) kernel on both Ubuntu Server and CentOS 8.
How to install Linux kernel 5.9 on Ubuntu Server (20.04)
Log in to your server and download the necessary .deb files with the following commands:
cd /tmp wget -c https://kernel.ubuntu.com/~kernel-ppa/mainline/v5.9/amd64/linux-headers-5.9.0-050900_5.9.0-050900.202010112230_all.deb wget -c https://kernel.ubuntu.com/~kernel-ppa/mainline/v5.9/amd64/linux-headers-5.9.0-050900-generic_5.9.0-050900.202010112230_amd64.deb wget -c https://kernel.ubuntu.com/~kernel-ppa/mainline/v5.9/amd64/linux-image-unsigned-5.9.0-050900-generic_5.9.0-050900.202010112230_amd64.deb wget -c https://kernel.ubuntu.com/~kernel-ppa/mainline/v5.9/amd64/linux-modules-5.9.0-050900-generic_5.9.0-050900.202010112230_amd64.deb
Once all of the files have downloaded, install them with the command:
sudo dpkg -i *.deb
When the installation completes, reboot, and you should be good to go. After logging back into the system, issue the command:
You should see that your Ubuntu server is now running kernel 5.9.
How to install Linux kernel 5.9 on CentOS 8
Now let’s do the same for CentOS 8. Log in to that server and first install the GPG key for the necessary repository with the following commands:
cat /etc/redhat-release uname -srv sudo rpm --import https://www.elrepo.org/RPM-GPG-KEY-elrepo.org
Install the repository with the command:
sudo dnf install https://www.elrepo.org/elrepo-release-8.0-2.el8.elrepo.noarch.rpm
Install the mainline (5.9) kernel with the command:
sudo dnf --enablerepo=elrepo-kernel install kernel-ml
Restart your CentOS machine. Once you’ve logged back in, issue the command:
You should see the Linux 5.9 kernel listed.
Congratulations! You’ve successfully installed the 5.9 Linux kernel on both Ubuntu Server and CentOS 8. Do this on testing machines or virtual machines to see if this latest kernel lives up to the performance boost hype it’s received.