Book Image

Linux Service Management Made Easy with systemd

4 (1)
Book Image

Linux Service Management Made Easy with systemd

4 (1)

Overview of this book

Linux Service Management Made Easy with systemd will provide you with an in-depth understanding of systemd, so that you can set up your servers securely and efficiently.This is a comprehensive guide for Linux administrators that will help you get the best of systemd, starting with an explanation of the fundamentals of systemd management.You’ll also learn how to edit and create your own systemd units, which will be particularly helpful if you need to create custom services or timers and add features or security to an existing service. Next, you'll find out how to analyze and fix boot-up challenges and set system parameters. An overview of cgroups that'll help you control system resource usage for both processes and users will also be covered, alongside a practical demonstration on how cgroups are structured, spotting the differences between cgroups Version 1 and 2, and how to set resource limits on both. Finally, you'll learn about the systemd way of performing time-keeping, networking, logging, and login management. You'll discover how to configure servers accurately and gather system information to analyze system security and performance. By the end of this Linux book, you’ll be able to efficiently manage all aspects of a server running the systemd init system.
Table of Contents (23 chapters)
1
Section 1: Using systemd
12
Section 2: Understanding cgroups
16
Section 3: Logging, Timekeeping, Networking, and Booting

Controlling memory usage

Let's start by having Vicky do something that will hog all of the system memory. As before, we'll use the stress-ng utility to simulate that, like this:

vicky@ubuntu2004:~$ stress-ng --brk 4

Wait a few moments, and you'll see some fairly ugly things in the top display:

Figure 12.9 – The top display for Vicky's memory usage

Yeah, only 98.9 bytes of free memory, and super-high load averages. In fact, after about 2 minutes or so, this virtual machine is completely unresponsive to any commands. Ouch!

Now, understand that I still have the 200% CPUQuota set for Vicky. So, CPU usage isn't the problem here. The load average is a representation of how many tasks are waiting to be serviced by the CPU. In the top part of the top display, as shown in Figure 12.9, the 53.51 that you see is the 1-minute average, 46.38 is the 5-minute average, and 25.00 is the 15-minute average. These load averages are...