Two years ago we announced the systemd project and began integrating it into the distributions. Since then it acquired a lively and large community, has been adopted by a variety of commercial and community distributions, and has gained wide acceptance in the desktop, server and embedded world. Today, you can buy devices and appliances built on systemd, and systemd is the basis of industry standards. It is time to look back on the first two years of systemd, analyze what we achieved, where our successes are and where our weaknesses. It's time to discuss where we want to take the project next and what to focus on.
This talk is intended for all technical folks, administrators and developers alike, as well as everybody else who is interested in the systemd project, its strenghs and the next steps.
Operating system-level virtualization allows running multiple isolated Linux containers with only one kernel. It has the smallest overhead and is used since a decade by the webhosting industry for virtual private servers (lately called cloud server).Linux-Vserver started in 2001, followed by Virtuozzo/OpenVZ with big kernel patches.In contrast LXC has the goal to bring everything into mainline kernel. It's included since 2.6.29 and most of the distributions support it. LXC makes use of kernel namespaces for isolation and uses cgroups (control groups) for resource limits. Best practices, pitfalls and how LXC can be used in HA environments with Pacemaker are shown. The new apparmor profile that makes the use of LXC more secure is also discussed.Audience: The presentation is intended for system architects, administrators and developers. The level of technical experience is moderate.
When you install Linux Server it is optimize for average workloads. With most servers you can gain much by optimizing performance. In this session you will learn how to optimize your server's performance by tuning kernel parameters, exclude unneeded system services and make some more tweaking.
This talk covers building a distributed database cluster and monitoring it. The tools covered will be HBase, Chef, OpenTSDB. If you are a large system administrator, or have been saddled with administering a Hadoop cluster, this talk will cover some of the finer points learned from building and administering multiple such clusters in the past. Much of the experience gained will be applicable in non-Chef and non-Hadoop environments (for example, if you deploy Cassandra via Puppet).
The Open Source community is now taking HA on Linux to the next step by building clusters of clusters that can recover from disasters that take out a whole site; allowing a reliability infrastructure that spans multiple cities or even continents. Linux provides even built-in support for asynchronous storage replication, and will further aid administrators by easing configuration replication and adjustment across multiple sites. This presentation will introduce the new software components, discuss deployment patterns, and present the current state, opportunities for extensions, and the future roadmap.
The target audience includes system architects, CTOs, and interested partners that which to contribute to the projects and collaborate in this area.
With an increasing security awareness among web and cloud developers, knowing how to secure your database from unauthorized or malicious access has become important. This talk explains the MySQL security model, pluggable authentication, new auditing features and rounds off with some pointers on how to securely integrate your database into your Linux web stack.
Server resources have exploded recently (# of cores, RAM size, I/O, ...). Previously seen only on MIPS or Itanium architectures, x86 has also recently seen machines with mainframe type capabilities.This presentation will detail the various approaches available on Linux to control and manage optimally these precious resources.We will cover the 3 main technologies:
Each time, use cases will be given to illustrate the best fit of each technology in a global portfolio. A cgroup demo will also be performed to show its capabilities. Audience expected: System Administrators and Technology Architects willing to optimize their server usage.