LCA2017 Sysadmin Miniconf Presentations
The abstracts for the presentations accepted for the Linux.Conf.Au 2017 Syadmin Miniconf are listed below. A draft presentation schedule is online now.
(Links sorted by presentation title; Abstracts below are sorted by first name of presenter.)
- Continuous Dashboarding - your devops airbag - Christopher Biggs
- Go for DevOps - Caskey L. Dickson
- Lightweight inventory management with Pallet Jack - Karl-Johan Karlsson
- Managing performance parameters through systemd - Sander van Vugt
- The Opposite of the Cloud - Tom Eastman
- Running production workloads in a programmable infrastructure - Alejandro Tesch
- Samba and the road to 100k users - Andrew Bartlett
- The School for Sysadmins Who Can’t Timesync Good and Wanna Learn To Do Other Stuff Good Too - Paul Gear
- The Sound of Silencing - Julien Goodwin
- Turtles all the way down - Thin LVM + KVM tips and Tricks - Steven Ellis
Infrastructure as Code is a hot topic today, however several IT managers and System Administrators often fail to see the benefits that it can provide to their day to day operations.
In this mini-conf session we will go in depth into two use cases that can leverage on OpenStack to address common operation challenges.
1) We will demo how to leverage HEAT and Ansible to automate the deployment of a two-tier load balanced environment within seconds that will host your enterprise applications. We will showcase as well the autoscaling capabilities by overloading our web farm until we hit the CPU threshold that will spawn another tomcat server.
2) "Data Cloning as a Service" will demonstrate how to automate the cloning of a 24 x 7 Mission critical Oracle database without impacting the end users.
Ansible is used to place an enterprise database in backup mode, take a snapshot of the volume containing the Database, clone the cinder volume and get it transferred to a development tenant who will automatically start a clone of the database for testing and QA.
About Alejandro Tesch:
Alex has been working with Open Source enterprise technologies for the better part of his 15 years IT career in companies like Hewlett-Packard Enterprise, Red Hat, IBM and Sun Microsystems.
He has extensive deployment experience on configuration management and automation of private cloud based on OpenStack and Docker microservices following devOps methodologies by leveraging on Jenkins and Github.
Alex is currently an APJ Cloud Consultant in the Helion Cloud team at Hewlett-Packard Enterprise, where he evangelizes the OpenSource side of the Helion portfolio (OpenStack / Cloud Foundry / Ceph).
He enjoys running automation workshops and seminars in the APJ region for cloud and docker adopters.
Samba4 (our Samba AD DC) is increasingly widly used and is becoming used at scale, and this talk will detail Samba's journey from 'wow, it works' to 'can we have 100,000 users please'.
Samba 4.5 brings significant improvement in our scale, to around 30,000 users, and this talk will look at how we got there, share some war stories along the way, and what we are doing to bring us up to the 100,000 size.
About Andrew Bartlett:
Andrew Bartlett is a Samba developer working for Catalyst in Wellington, NZ
With over 15 years as a Samba developer, and a long-standing member of the international Samba Team, Andrew now leads a team at Catalyst with a near-exclusive focus on the Samba Active Directory DC, otherwise known as Samba4.
Go is an excellent open source multiplatform language for doing cross-platform development which is ideal for ops tools. If you target multi-platform, multi-architecture environments, come learn how Go enables you to write once and run anywhere. No more need for kludgy shell scripts!
This talk will cover the built in features of Golang that make it possible to write your code on one platform and seamlessly deliver it to many different operating systems and architectures. As a devops engineer I have tools and services that need to support multiple operating systems running on different CPU architectures.
You will learn the specific ways to generate *bsd-386 binaries on Windows-x64, android-arm binaries on Linux-ppc and all combinations in between. I'll also be covering coding techniques to keep this manageable and libraries that support you in things like file system navigation.
About Caskey L. Dickson:
Caskey L. Dickson is a Site Reliability Engineer at Microsoft where he is part of the leadership team reinventing operations at Azure. Before that he was at Google where he worked as an SRE/SWE writing and maintaining monitoring services that operate at "Google scale" as well as business intelligence pipelines. He has worked in online services since 1995 when he turned up his first web server, discovered Linux at version 0.95, and has been online ever since. Before working at Google, he was a senior developer at Symantec, wrote software for various Internet startups such as CitySearch and CarsDirect, ran a consulting company, and even taught undergraduate and graduate computer science at Loyola Marymount University. He has a B.S. in Computer Science, a Masters in Systems Engineering, and an M.B.A from Loyola Marymount.
The concept of "Continuous Delivery" has been paraphrased as "Move fast and break things", that is automate your testing and deployment to remove as many delays as possible between developer and production, and be prepared to react rapidly when something goes wrong.
The trick is knowing whether something has gone wrong.
Most businesses have analytics, monitoring and dashboards.
In fact, if you look really hard, most businesses have a dozen or more different silos of monitoring data. The problem is that it's difficult to get a business-wide overview of all this data.
Node-RED is a visual data-flow tool that has a rich collection of interfaces to other services, and is capable of talking to most of these silos in order to provide synergistic insights, and real time alerts.
In combination with mobile dashboard tools like Blynk, this allows targeted mobile dashboard applications to be built and deployed in minutes, for one user or a thousand.
The attendee will learn the capabilities of Node-RED and Blynk, see some examples of dashboard apps, and receive practical advice on how to get started with their own data.
With "dashboard driven development" your process becomes first build a dashboard to monitor some element, second make a change, third verify that the change had the desired effect.
Every time you make a change to your operating environment, you'll be able to monitor vital statistics from your pocket, in real time.
The collection of dashboards you've built over time will give you increased confidence whenever you make a change that your business' key performance metrics are where you expect them to be.
About Christopher Biggs:
Christopher has been into Unix since 1990 and was there at the birth of 386BSD and Linux. He was founding vice president of HUMBUG, the Brisbane open systems user group, and was head of development at a multinational online retailer, stewarding development culture and toolset, and convening a series of lectures and workshops for developers there.
In his spare time he builds and blogs robots, and adds to the Internet of Things.
Having automatic monitoring alerting us when there's problems with our systems is a core part of ensuring reliable systems, however this can lead to alert fatigue.
This talk will cover using silencing, grouping and ihibition to reduce alert load, and improve the signal to noise ratio of remaining alerts.
About Julien Goodwin:
Julien is a Site Reliablility Engineer at Google, working day-to-day maintaining one of the worlds largest IP networks. In the past he has worked as a (primarily Linux) Sys Admin on educational networks, and in small businesses. He has a long history working with the Australian FOSS community and was a member of the LCA2008 team, and has spoken at several previous LCA's and Sysadmin Miniconfs, most recently in 2015.
Inventory management for systems administrators is crucial for day-to-day operations. When installing a new machine, its MAC address needs to be registered with the DHCP server. While it runs, the configuration management system needs to know what role it has so it can install the correct software. When it breaks, you need to know which rack and switch port to find it in, its serial number and when its service contract expires.
We present Pallet Jack, a systems administrator-friendly inventory database.
Its back-end store consists of a directory tree with YAML fragments, which means you can insert, version and extract information using standard text processing tools. Inheritance and automatic text processing eliminates the need to store duplicated information. Finally, a simple but powerful Ruby library helps with generating configuration files from the database.
About Karl-Johan Karlsson:
Karl-Johan has been a Linux systems administrator for twenty years, but still finds new and exciting ways for systems to break. Having installed systems in South Africa, he now makes sure to include pest control when writing site requirements. He currently builds flight simulators for a company that doesn't agree with him that everything should run bleeding edge Gentoo.
For relaxation, he flies light aircraft with as few computers in them as possible.
Network Time Protocol (NTP) is a widely deployed but oft-misunderstood Internet standard for time synchronisation. This presentation will cover why you should (and probably already do) use NTP, how it works, and an overview of configuration, troubleshooting, and monitoring. The focus will be on practical implementation in enterprise, cloud, and personal environments rather than history, theoretical grounding, or environments requiring extreme precision.
Topics covered will include (time permitting):
- overview of the NTP algorithms, strata
- installation & configuration with
- how to use ntpq for troubleshooting
- overview of monitoring tools
- how to use and participate in the NTP pool & comments on scale
- common myths & misconceptions, and best practices for avoiding them
- my experiences in building a sub-US$100 stratum 1 server
For more on the history and theoretical background of NTP, see Julien Goodwin's LCA2011 talk, "A Brief History of Time Synchronisation". Overlap with Julien's talk will be kept to an absolute minimum, although I will dispute/update some of his data.
The practical examples & demo will use the NTP reference implementation on Ubuntu 16.04 LTS, but the skills learned will be applicable to all Linux distributions, and the principles will likely apply to other NTP implementations and operating systems as well. The content on which this talk will be based is appearing in a series of blog posts, starting here.
About Paul Gear:
Paul first encountered Linux in 1991, when trying to do something more useful with his 386SX Minix system, and has been hooked on it ever since. He has brought a passion for Free Software to various roles in development, systems administration, network engineering, and IT management since then. In 2013, Paul started the LibreNMS project, a network monitoring system focused on ease of use and community participation. Paul works as a Systems Reliability Engineer in Canonical's global Information Systems team.
In this talk you¹ll learn how to manage performance parameters through systemd. You¹ll discover how systemd splits the Linux operating system into 3 different slices - and how that may lead to unexpected situations in some cases. You¹ll also learn how to modify a unit file to ensure that your services will always be sure that the resources they need will be available.
About Sander van Vugt:
Sander van Vugt is an author and Linux instructor who has been working with Linux since the very beginning. Some of his works include the book "RHCSA/RHCE 7 Cert Guide" and his new video course "Linux under the hood". Sander likes digging in to the possibilities of the operating system, and particularly likes getting beyond what is provided by command line interfaces. San
Tips and tricks for both operators and developers to get the most out of ThinLVM and KVM when building demo environments, or optimising real world business solutions.
Topics covered will include
- Thin LVM
- Thin VMs on Thin LVM
- TRIM for SSDs + Thin VMs on Thin LVM
- TRIM for SSDs + Thin VMs on Thin LVM + Luks Encryption
- KVM performance tips and tricks on ....
Basically "Turtles all the way down.."
About Steven Ellis:
Steve's day job with Red Hat is to persuade organisations to spend their IT budgets on Open Source rather than traditional proprietary technologies. With over 20 years experience of Open Source from development to enterprise architecture his passion for Open Source helped bring linux.conf.au to Auckland NZ back in 2015.
In his spare time he still hacks on MythTV and debugging random new bits of hardware that really should know better.
Steven gives regular talks on FOSS as part of his day job at Red Hat, including technology briefings and community meetups.
In addition he's presented at linux.conf.au 2008, OSDC 2008 / 2009 / 2013, Linux World, OSCON, OpenStack Summit, and a regular presenter at the linux.conf.au Sys Admin miniconf.
Cloud computing. It's referred to -- humourously, or bitterly depending on whether you're talking to a 'Dev' or an 'Ops' respectively -- as "Somebody Else's Computer".
If "The Cloud" is hosting your infrastructure on "somebody else's computer", then the opposite of the cloud is when you're faced with putting someone else's infrastructure on your own internal network.
And so here I am, tasked with building virtual machine appliance images to safely and securely run behind customer firewalls of unknown configuration.
That... is an exercise in trust. Trust that is getting harder to come by thanks to the embarassing security record of "Internet Of Things" devices.
This is a talk about DevOps practice, infrastructure architecture, command & control, VPNs, Docker, Debian, Saltstack, Packer, Terraform, careful orchestration, and serious paranoia.
About Tom Eastman:
Tom is an open source technologist, Python developer, trainer, devops/security consultant, and senior software systems engineer at Koordinates Limited. He believes your two crucial metrics for measuring code-quality should be (a) "Will the person who inherits my code be glad that I wrote it this way?" and (b) "Will the person who attacks my code be annoyed that I wrote it this way?"