Proposed Sessions

Varnish In Action

ThijsFeryn — Wed, 19 Dec 2012 17:07:46 +0000

Having trouble scaling your website? Is that new marketing campaign killing your performance? Are you using a frameork or a CMS that causes too much overhead? Give Varnish a try !

Varnish is a reverse proxy that serves cached pages to your visitors instead of hitting the backend servers all the time. It's an insanely powerful tool that will rescue your site from getting crushed and is considered to be an industry standard.

Cache is stored in RAM and there's a domain specific language that allows you to hook into your caching logic. There are tools and rules and it's essentially your primary tool for when you're site is "at load".

Did I mention it works great with Drupal?

Slides:

varnish_in_action.txt

Schedule info

Status:

Proposed

Session Info

Speaker(s):

ThijsFeryn

Track:

DevOps

Featured:

Not Featured

Experience level:

Intermediate

Reverse caching proxies: Varnish or Nginx?

ThijsFeryn — Wed, 19 Dec 2012 17:16:29 +0000

In a world full of scalability challenges, a reverse caching proxy has become an essential tool to protect your backend servers. This talk is dedicated to that and compares Varnish to Nginx as the 2 tools for the job. We'll be talking about setting the up, playing around with config, making them serve cached Drupal or PHP content and do some advanced tricks such as ESI

Schedule info

Status:

Proposed

Session Info

Speaker(s):

ThijsFeryn

Track:

DevOps

Featured:

Not Featured

Experience level:

Intermediate

The Proper Care and Feeding of your MySQL Database

davidstokes — Mon, 07 Jan 2013 21:25:45 +0000

MySQL is the database for most Drupal sites. But what do you need to do to keep your database happy? What signs do you watch for to make sure it is healthy? How often do you have to backup data? This session will cover what MySQL does with your Drupal data, the five things you must do to keep the database (and your boss) happy, simple scaling tricks that do not require massive hardware investements, and what new features in MySQL 5.6 that you can exploit to make your Drupal sites faster.

Schedule info

Status:

Proposed

Session Info

Speaker(s):

davidstokes

Track:

DevOps

Experience level:

Beginner

Building a huge content sharing network using Domain Access

aburrows — Wed, 09 Jan 2013 14:06:14 +0000

So you have multiple websites and you want to share content and users across multiple domains? But wait you want it to be simple and managed from one domain? Well you could use multisite, but that can be tedious and cause lots of issues especially with a huge amount of sites and databases required, different content types, client changing their mind non-stop, new funcitonality arising at each step. Once again Toyota UK dealer websites required this and the previous methodology behind it was to use multisite and XMLRPC, but this wasnt successful and required a lot of custom module development. I'll explain the pros and cons of multisite and domain access, why domain access was used, the cool integration with views and content types and basic permissions of domain access.

Schedule info

Status:

Proposed

Session Info

Speaker(s):

aburrows

Track:

DevOps

Experience level:

Intermediate

Scalability and high-performance: Lessons learned from hosting the largest Drupal sites on Acquia Cloud

ricardoamaro — Thu, 10 Jan 2013 22:19:57 +0000

The world of Drupal grows faster in each passing day!
Large companies, especially in media, entertainment and services, seek for solutions on rapid and reliable development for the web. There Drupal plays a major role.

However, it remains a big question: "How and where to scale, with stability, huge sites, made in Drupal?"
The Acquia Cloud mission is to address these issues and grow Drupal as the chosen platform worldwide. From standalone EC2 Cloud servers, to Multi-region worldwide setups.

There are many moving parts when it comes to hosting any website - database tuning, caching layers, Drupal patches and settings, monitoring, backup, staging and more.

In this session we will share the steps taken and lessons learned, while hosting the world's largest sites. Automation, configurations, optimizations, systems engineering, operations ... Crucial factors hard learned!

- What tools to use for scaling Drupal?
- How to pinpoint and solve performance issues?
- Key Customers with Sites needing an audit to grow fast.
- Helping customers understand your systems.
- Maintaining an hosting codebase with highly dynamic performance.
- Monitoring thousands of servers efficiently.
- Have a contingency plan and invest on security.
- Plan for failure.

Schedule info

Status:

Proposed

Session Info

Speaker(s):

ricardoamaro

Track:

DevOps

Experience level:

Intermediate

Drupal + PAAS = Drupal App

rlnorthcutt — Thu, 10 Jan 2013 23:30:24 +0000

While "cloud" may be a nifty new buzzword, the reality is not quite so simple. There are many ways and methods for service Drupal sites "in the cloud". In this session, we will look at several PAAS offerings, how they work, and even how to set up Drupal in a site.

We will compare and contrast:
- Pantheon
- Acquia Dev
- Appfog
- Orchestra PHP
- Openshift

They all have different pricing structures and feature sets, but all are really just convenient front ends to host your Drupal (or other PHP site) in the cloud. Most also require some type of external file storage (like S3)

For fun, we will also create a hosted Drupal site on Openshift, so sign up for a free Openshift account and follow along in the session.

This session is designed for intermediate users that have basic knowledge of :
- Git
- SSH
- Devops in general

NOTE: PAAS means Platform as a service

Schedule info

Status:

Proposed

Session Info

Speaker(s):

rlnorthcutt

Track:

DevOps

Experience level:

Intermediate

Baby Steps: Get Your DevOps Process Started By The End of This Session

ldpm — Sat, 26 Jan 2013 06:00:20 +0000

Alternate title: "Okay, we've installed Jenkins. Now what?"

DevOps is in the sweet spot for many medium size development firms, but it can be a tougher fit for others. When management is used to doing things a certain way, they can perceive the barriers to entry and risks of "merging" the development and operations teams to be too great to justify the potential rewards. Smaller teams or firms might not want to invest the time to get a big Java-style build server configured, especially when they don't host production themselves.

But DevOps doesn't have to be all or nothing and doesn't need new squares on the Org Chart to see benefits. In this session, we will examine low-risk changes you can make to the way you do things now that will start you down the DevOps path:

Writing more testable code

Moving your unit and functional tests outside of Drupal

Using commit hooks

Getting your project managers involved

Integrating with a third party Operations Department

You should be able to see at minimum the following benefits:

Better code coverage with automated tests (or at least better awareness of gaps of coverage)

Fewer bugs making it to staging, let alone production

Faster turnaround for feature requests

If you see those benefits, you might decide to drink more of the DevOps Kool-Aid; if not, you at least know what you're missing.

Schedule info

Status:

Proposed

Session Info

Speaker(s):

ldpm

Track:

DevOps

Experience level:

Beginner

Linux system security tunables

keescook — Mon, 28 Jan 2013 21:35:09 +0000

Click here to watch Linux system security tunables.

While regular web hosting best-practices should always be in place to make your site defensive against attack, there are even more things you can do to improve your host's "post intrusion" stance. A layered set of security precautions can discourage an attacker's attempts at privilege escalation.

The first line of defense is privilege separation. Beyond keeping a system's administrative users separate from web server users, it's important to maintain strict control of any SSH keys or similar cross-host access mechanisms. Two factor authentication is surprisingly easy to implement. And in cases where there are clear lines between sites or services, adding mandatory access controls can help strengthen the existing boundaries.

To frustrate an attacker's information gathering and potential vulnerability exploitation, there are a lot of system tunables that can be trivially enabled. These span the network, filesystem, virtual memory, and debugging facilities of the kernel. Additionally, more can be done to protect the kernel from userspace. While it may look pointless to keep root user access separate from kernel-level access, it can play a significant role in hardening a system, especially those with external kernels as seen with some VPS providers.

Schedule info

Status:

Accepted

Time slot:

Tuesday, May 21 - 02:00pm-03:00pm

Room:

OR 201 - Phase2

Session Info

Speaker(s):

keescook

Track:

DevOps

Featured:

Featured

Experience level:

Intermediate

Introduction to Drupal DevOps

cyberswat — Mon, 28 Jan 2013 23:22:54 +0000

In case you missed the news, Drupal is growing. As Drupal grows it continues to evolve and become a critical component of mission critical, high performance systems. Development teams are growing and becoming more geographically dispersed. Systems needs are increasing as Drupal becomes more capable. The days of using a single LAMP machine to run a site were there for you to learn but do not apply in the enterprise world. I work regularly on complex Drupal sites that have as many systems communicating with each other and working together. It's no longer uncommon to manage multiple datasources (MySQL, SQL, MongoDb, Reddis, text files on s3, etc.) or multiple server types (Varnish, Memcache, NGiNX, Lucene, etc.) in addition to simply writing the code that got us here in the first place.

With so many facets adding to the complexity of an overall system it is easy to become overwhelmed. How many times have you deployed something only to find out it works great on your machine but fails miserably on the production server? What branching strategy should your developers use? How do you guarantee a stable environment that lets developers and product people innovate? What kind of a person can connect the dots between developers, infrastructure and business and why is this an important new field?

In this session I will introduce you to the world of DevOps and will walk you what it takes to successfully manage and deploy complex Drupal systems to a heavily trafficked and highly performant site. I will introduce you to some technologies you may have heard of, and I'll outline strategies that work while that minimizing the impact on your teams.

Schedule info

Status:

Proposed

Session Info

Speaker(s):

cyberswat

Track:

DevOps

Experience level:

Beginner

Bootstrapping Solr search clusters and maintaining them

Nick_vh — Wed, 30 Jan 2013 16:38:03 +0000

At Acquia we have the Acquia Search service. This service allows people to connect with Apache Solr (A search application) and also provides high availability and high scalability.

This session will contain :
- Basic understanding of Solr and examples of use cases where the most obvious solution is not always the best solution. We'll try to teach you to look further and make smarter decisions.
- How we use Ruby scripts and Nagios to monitor critical components such as query time, memory, etc...
- What is an optimal server for Solr, and how to build it using Puppet
- Request handling with Nginx to help our load balancing process
- How this scales for up to +3000 Solr cores, using a Master and Slave, even a repeater.
- The role of Solr 3 vs Solr 4, and some small discussion perhaps about elastic search
- How we fulfills our needs in terms of spawning new solr cores for customers automatically.
- How we created authentication for Apache Solr while keeping its speed and keeping it simple to use.

This process took time to build so I hope this session can bring developers up to speed with the modern techniques and hopefully it can create some itches here and there so people start exploring. Setting up your own Solr Server is quite easy, but you don't want to wake up and figure out the server needs more capacity, or even worse, it is down. You certainly don't want to repeat the whole process to build an exact copy of an existing system for another client.

If you are interested in the internals of Solr and how to manage this then this session has been build for you.
If you are interested in Solr from a developer point of view, you are certainly also welcome.
We'll give hints and tips to speed up Solr and to make better use of its cache system.

We will not go in depth module wise, a BOF will be appropriate for the more in-depth Solr/Search questions, developments and conversations.

Slides:

2012-08-solr-devops-08-20-15gmt.key

Schedule info

Status:

Proposed

Session Info

Speaker(s):

pwolanin

Nick_vh

Track:

DevOps

Experience level:

Intermediate