LCA2019 Sysadmin Miniconf Presentations
The abstracts for the presentations accepted for the Linux.Conf.Au 2019 Syadmin Miniconf are listed below.
A draft presentation schedule will be created in December 2018.
(Links sorted by presentation title; Abstracts below are sorted by first name of the first presenter.)
- Adventures in highish availability - Peter Chubb
- Dirty Samba hacks to make your life easier - Kirin van der Veer
- Distributed storage is easier now: usability from Ceph Luminous to Nautilus - Tim Serong
- Keeping the Balance: loadbalancing demystified - Murali Suriar
- Petabytes Data Migration and Load Balancing with Football+MARS - Thomas Schoebel-Theuer
- Prometheus - For both big and little people - Simon Lyall
- Python++ – Bringing your code to the next level - Jan Groth
- Samba for the 100,000 user enterprise: are we there yet? - Andrew Bartlett
- Survival guide for women in IT - Anna Fiofilova
- The Free Software Mirror Group - Brad Cowie and James Forman
- Toolmaking - Julien Goodwin
- Using containers (but avoiding Docker) for HPC workloads - Eric Burgueño
For a number of years Andrew Bartlett has given presentations along the lines of 'the road to 100k users', and this so in part, this is another update on that long road.
But it is also a milestone, because in the past year Samba has successfully added 124,000 users to the DB in a 4-hour benchmark, re-targeted to LMDB and successfully stood up a second replicated domain controller at the 100,000 user scale.
Beyond scale, Andrew Bartlett will also update the audience on the many other new features coming up in Samba 4.10.
About Andrew Bartlett:
Andrew Bartlett is a Samba developer and has been at the forefront of Samba's Active Directory DC effort since the Samba4 effort started in 2003.
Andrew now leads a Samba development team for Catalyst in Wellington, New Zealand
Working in the IT industry today is hard, but it is even harder if you are a woman. There is still lots of “old school thinking" that women face daily. This talk is a survival guide based on real-life stories and different challenges from women of different ages, cultural backgrounds and roles in the Australian IT industry.
Like any survival guide, this one provides you with the essential information to help you identify and overcome the most frequently encountered hazards. Each chapter contains useful tips, instructions and practical advice on a particular issue so you can implement the skills and techniques even under the most stressful circumstances. From the hiring process to promotions and corporate events, you'll have the tools to survive.
You will learn these skills and more:
- Assess your situation and prioritize your needs;
- Surviving techniques for the hiring process;
- Assemble your own custom emergency kit with essentials tools;
- Manage extreme work conditions and overtime;
- Survive corporate parties and drinks;
- Build trust network and create allies;
- Identify your enemies and their habits.
Preparation is the key. If you are starting your career in IT or navigating through it - this guide is for you.
About Anna Fiofilova:
I was born in the USSR, a country that doesn't exist any more. After it fell apart, I grew up in Ukraine, close to the Russian border.
I was reasonably good at mathematics and had learned basic programming at school, so in 2002 I enrolled at my local University to study "Computer Systems and Networks".
I started my career in IT after finishing my second year in Uni; I got my first real, paid job in 2005 while studying my bachelor's degree. I stayed in the University for two more years to get my masters degree while working full time. I was changing jobs roughly every two years, to accrue knowledge and to stay up to date with the latest shiny tech.
I moved in Melbourne in 2012 (just before the Higgs boson was discovered) and was lucky enough to join ThoughWorks. At ThoughtWorks I found out about gender diversity, what 'agile' really means, and had the chance to meet many smart people. Working at ThoughtWorks was a big shift for me in terms of working environment and culture.
After 6 month in ThoughtWorks I joined REA group as senior software engineer and still enjoying working there. Now I am an Australian citizen and proud to be part of its diverse culture.
A lot of the newer free software projects use either cloud providers or a small set of mirrors located in Europe or the USA to provide access to their software and packages. New Zealand has a unique situation where we have ubiquitous access to fast fibre in urban centres and high latency low bandwidth to International destinations. In order to improve access to popular and new free software projects a number of local sysadmins from different companies formed “The Free Software Mirror Group” (FSMG).
We will take this opportunity to introduce the community to FSMG and propose the ways we want to help the community in future. This talk will cover what happened in FSMG’s first year of operation, from how we started the project to how we operate the project day to day.
FSMG are currently official mirrors in New Zealand for Debian, Kali, Raspbian, CentOS, Arch Linux, OpenBSD and other projects.
About Brad Cowie:
Brad is a member of the WAND Network Research Group at the University of Waikato. He is also a core member of the FAUCET project which develops an open source OpenFlow controller for enterprise networks. Utilising his years of experience deploying servers and networks he deploys FAUCET around the world doing SDN deployments with FAUCET.
About James Forman:
James is a Linux Sysadmin at Catalyst IT in Wellington where he works on building things that can survive earthquakes. He has obsessions with monitoring, automation, backups and making sure people don’t get woken up.
James is a maintainer of New Zealand’s Free Software Mirror Group.
Containers are the old hot thing. But their adoption in all realms of IT is far from widespread. Most containers we see in the wild today are created to host daemon-type processes and scale them massively, but traditional HPC workloads are different and so HPC shops continue to have some aversion to them.
In this short talk I will discuss some of the challenges that Docker introduces when it comes to using containers in a multi-tenant HPC cluster, and what are some of the possible solutions. We will also explore the relationship between containers and reproducibility in Computational Science.
About Eric Burgueño:
Eric Burgueño is a law school graduate that prefers to work in IT as a SysAdmin. After bouncing around "big name" companies in his home country, he's settled for a more peaceful but challenging kiwi life.
He currently makes the world a better place by helping the scientists that take care of the "making the world a better place" part.
Python is a great language for DevOps tasks. It’s easy to use for automation and offers an end-to-end range of tooling for managing infrastructure on-premise and in the cloud. Scripts are quickly implemented and new features easily rolled out. But what if complexity grows and all of the sudden you find yourself in a complete mess? How do you add a new feature or fix a bug in a script that you struggle to understand because it was written months ago? And do you sometimes see someone else’s code that you like but can't always put your finger on the magic ingredient? This talk is aimed at you if you are reasonably confident reading Python code and want to discover and improve beyond the basics. I'll provide you with ideas and suggestions that will help you stay on top of your coding and will bring it to the next level: Clean code which is easier to understand, more functional, testable and beautiful.
About Jan Groth:
Jan is happy scripting and automating in Python as a Cloud Engineer at Versent/Sydney. Prior to moving into DevOps he has been a Java Developer for almost 15 years.
From simple shell pipelines through more advanced query & workflow systems basic toolmaking is one of the most valuable ways to increase the efficiency of operations work.
In this presentation we look at toolmaking in other fields, and how to best scale toolmaking work to keep the efficiency wins going.
About Julien Goodwin:
Julien is a Site Reliability Engineer working on some of the largest networks on the planet. He has spoken many times at the LCA SysAdmin miniconf, as well as at OSDC and other conferences in Australia, New Zealand & the United States. He was part of the organising team for linux.conf.au 2008 in Melbourne.
Thanks to the proliferation of IoT devices Samba is everywhere.
But what if you want to use it the "normal" way? You know, as a DC and file server for a large number of users?
What happens when you need to parse the directory for a list of disabled users and take action? Or have regular helpdesk guys deal with locked files?
For regularly running scripts what returns queries faster; pdbedit or ldapsearch?
Perhaps you'd like to setup a linux based hotdesk that will automatically RDP into whatever workstation a user happens to be logged into elsewhere on campus. How to you remotely (and securely) request that information from Samba?
How about a website that acts as a password reset portal? (don't forget your input validation!)
During this talk I will outline and demo a number of scripts and hacks used to make Samba work for our organization (including links to source code).
About Kirin van der Veer:
Kirin van der Veer has built and Sysadmin'd networks for organizations as disparate as Greenpeace in Amsterdam and Military deployments in the middle east.
He occasionally files bug reports for Samba and other open source projects.
Currently he is the Senior (read only) Linux systems administrator for Planet Innovation in Melbourne.
In his spare time he once commissioned a developer to create a fork of Gnome terminal, simulating an accurate recreation of 1995 era VGA text modes in order to play ADOM the exact way he remembered.
Can you explain the entire path that an IP packet takes from your users to your binary? What about a web request? Do you understand the tradeoffs that different kinds of load balancing techniques make? If not, this talk is for you.
Load balancing is hard, and it is made up of many disparate technologies. It cuts across network, transport and application layers. We'll describe different flavours of load balancing (network, naming, application) and how they work.
We will then discuss example use cases, and which load balancing approaches are most appropriate in each case. We'll also relate these to several design patterns for high-availability services that have developed over the years. Finally, we'll relate the techniques we've been discussing to well-known open source technologies and to the major cloud load balancing services.
You will come away with:
- An understanding of various load balancing techniques, and some of their high points and pitfalls.
- Some criteria by which to evaluate different load balancing technologies.
- A framework of heuristics to use when deciding what load balancing pattern(s) to use with your service.
About Murali Suriar:
Murali Suriar is a lapsed computer science graduate, turned network engineer, turned SRE. Currently working at Google running cluster filesystem and locking services. Left Google to get on a boat. Got bored and came back.
I manage a small farm of servers and network for continuous integration and development, supporting around 50 users. We recently retired about a dozen servers, and have instead used containers and virtual machines on a pair of really big servers.
Given some excess capacity in the new machines, I decided to try to set up replication and failover, so I can bring one machine down for maintenance, and people won't notice (much). Although there are off-the-shelf tools (like Pacemaker), they didn't seem applicable --- so we rolled out own.
In hindsight this may have been a mistake.
In this talk, I'll be talking about all the things that went wrong.
About Peter Chubb:
Peter has been a long-term contributer to open source (his first patch was to international iSpell in around 1990 to enable Australian spelling rules), to Unix (he worked on the Unix kernel for SGI, Fujitsu, and for AT&T Bell Labs while at Softway Pty Ltd), and to Linux systems software (kernel and low-level software like u-boot and qemu). He has spoken at many LCA events.
Peter has never used a Windows operating system except when forced to, and then for only a short time.
Prometheus is a monitoring and alerting system that has gained wide popularity in the last few years. It is especially popular on cloud and kubernetes clusters.
This talk will running the software in both small (a few machines) and larger sites covering both similarities and differences between them.
About Simon Lyall:
Simon lives in Auckland, New Zealand and has worked as a Linux Sysadmin for over 20 years. He currently works for for a large New Zealand company wrangling Kubernetes and CI/CD Pipelines.
He has been attending Linux.conf.au since 2004 and has co-running 12 Sysadmin Miniconf's at LCA since 2006. He also maintains the "Unofficial Linux.conf.au Guide".
MARS is used at 1&1 IONOS for hardware lifecycle and load balancing of a few thousands of stateful LXC containers, in addition to its traditional long-distance data replication capabilities through network bottlenecks.
The talk explains some background, the Football toolset, and discusses experiences gained from a major migration project to newer hardware, increasing density, thus saving power costs of some millions of € per year, while improving customer performance.
About Thomas Schoebel-Theuer:
Dr. Thomas Schöbel-Theuer is the inventor of the Dentry Cache of the Linux Kernel. He has also implemented some research Operating System. Currently he is working on long-distance data replication of some petabytes.
Distributed storage is *still* complicated, but the Ceph project has put significant effort into making life easier for administrators and users over the past year.
Following on from Sage Weil's LCA 2018 presentation "Making distributed storage easy: usability in Ceph Luminous and beyond", this talk covers the changes we have made in the meantime, including major enhancements to the in-tree Ceph dashboard, centralised configuration management, placement group merging, and the beginnings of a new orchestrator module to tie in with Kubernetes, Salt and Ansible.
The Ceph Mimic release in June 2018 includes much of the above work, but we still have plenty more to do for the upcoming Ceph Nautilus release and beyond.
About Tim Serong:
Tim spends most of his time working for SUSE, hacking on Ceph and related technologies. He has spoken about high availability and distributed storage at several previous LCAs. In his spare time he wrangles pigs, chickens, sheep and ducks, and was declared by one colleague "teammate most likely to survive the zombie apocalypse".