Cascadia IT Conference Track 2

 
 


10:30 - 11:00   break


11:00 - 11:30   Talk: Chef: A Quick Look at Central State Management

Christian "Ian" Parades


There are several configuration management systems to choose from, such as Puppet, Bcfg2, cfengine, and Chef.  They all purport to fulfill the same purpose: provide a language to describe your set of machines and a desired system state that is OS/architecture agnostic, and then execute commands that converge the systems to your desired system state.  Many of these configuration management systems provide features that differentiate themselves from other configuration management systems; Chef's built-in ability to store system states in a central place and use them for converging other machines' system states is its differentiating feature from all of the others.  We will be looking at how this can be useful with a few real world examples, as well as using the `knife` command line tool for interacting with chef-server.


11:30 - 12:00   Talk: Configuration Management with Puppet

Garrett Honeycutt


In this short talk I cover why configuration management is important, how Puppet works, what the code looks like it, how it fits into your architecture, and how it differs from CFengine and Chef.


12:00 - 12:30   Talk: DevOps with Cfengine

Aleksey Tsalolikhin


This talk demonstrate the DevOps concepts of "Infrastructure as Code" and "production servers running without sshd" by using Cfengine alone to deploy, configure and integrate a multi-tier Web app in a cloud.


12:30 - 14:00   Lunch - on your own


Find a group of people and head off to any of the interesting restaurants in the area.  Large group or small, it’s a great opportunity to network!


14:00 - 14:30   Talk: Recovering From A Damaged PostgreSQL Cluster

Robert Bernier


How to recover a postgres cluster that has been damaged by Hardware failure.


This presentation guides the audience through the process of recovering a severely damaged cluster. This introductory talk summarizes the recovery process and the principles behind it.


He has made his slides available here: cascadia_postgres_rbernier.pdf



14:30 - 15:30   Invited Talk: A Few Thoughts on Uptime:  Resisting Entropy

Stuart Kendrick


This paper was born from my noticing that many small errors contribute to a service disruption.  Individually, these errors each seem preventable to me, using obvious techniques: clean up messes, fix what we know is broken, keep learning. Why is it, then, that we keep making these mistakes and keep knocking out services? I haven't answered that question -- but this paper offers snapshots from my search.  Bring your experiences, challenge my ideas, offer your insights.  http://www.skendric.com/philosophy/A-Few-Thoughts-on-Uptime.pdf



15:30 - 16:00   break



16:00 - 16:30   Talk: Bootable Cluster CD

Skylar Thompson


A common problem in computational science education, particularly at the high school and undergraduate levels, is that educators are unable to provide an environment to teach parallel computation. The Bootable Cluster CD (BCCD) provides a turn--key software tool to setting up a computational cluster, configuring everything from the network environment to MPI within minutes. This presentation would be suitable for educators and system administrators who support compute clusters in K-12 or undergraduate settings.



16:30 - 17:00   Talk: Deploying your Application to Cloud: Lessons Learned from Five Different Applications.

Brian Connolly


Over the past 2 years, I have been asked to migrate 5 different applications "to the cloud".  Three of the five applications were successfully migrated. For the other two applications, the migration to the cloud did not make sense for either performance or financial reasons. In this talk, I will review the types of clouds that are available. Discuss how to evaluate if an application is "ready for the cloud" and guidelines for how to move an application from your datacenter or traditional hosting company to the cloud. Then using each of the five applications as examples, I will discuss why there successfully migrated or not and the technical lessons learned during each migration.



17:00 - 17:30   Lightning Talks


Coming off its success at LISA 2009 and LISA 2010!  A series of five-minute talks where enthusiasm is paramount and almost anything that might be of interest to sysadmins is ontopic. Since we're lifting the idea in-toto, why not use their description too?  See http://www.allgoodbits.org/other/lt.html for details.

 

Saturday, March 12 — Track 2:

Regent Room