Jimmy Schementi

Get Your Infrastructure Right

Tue, 30 Dec 2014 00:00:00 +0000

Brooklyn J S

This is a write-up of a talk I gave at BrooklynJS on 12/18/2014,
titled “Build apps, not infrastructure”.

Use arrow keys or swipe to control slides | Open slides in a new window

While most of us love building apps, we’d also love it if people used them, and if we’re lucky, lots of people. Problem is, then we really need to care about how our app runs and scales, and it’s best to get it figured out before that moment of fame arrives. Let’s look at a few of the tools and services that you can use to manage your infrastructure, so you can spend more time building your apps, and less dealing with keeping your app running.

Have you ever...

Changed code

directly in

production?

Yes, we all have changed code directly in production (and you’re lying if you say you haven’t). We do this despite knowning that it’s horrible:

How many times have you found someone changed production directly and forgot to check it in?
Or made [what you thought was] a simple code change to fix a specific problem, but instead it revealed further problems.
Or users seeing intermittent brokenness while you make changes.
Or how do you know it works if you haven’t written tests for it?
Or! OR!

Heck, a company even has the person editing code in production don a pink sombrero to bring awareness to “Cowboy Coding”:

Our reasons for making changes directly in production are usually good. Sometimes they’re not. Regardless, we need to do this because our deployment processes aren’t automated enough to safely do this.

Think about it - if we could quickly write a fix in a clone of production, test it, and push it to production in seconds, wouldn’t we do that instead?

It’s a lot of work to fully automate deployments, so most people have a fairly manual deployment process. Which is crazy, considering we’re programmers - all we do is automate.

Application (feature)

vs.

Infrastructure (chore)

This happens mainly because we prioritize features, or enhancements to our applications, over chores, which can be enhancements to our infrastructure. This distinction comes from Scrum, where you’re rewarded for completing features, as those “add business value”…

Complete and utter bullshit…

Prioritizing

infrastructure

makes it easier to

ship features

A better infrastructure makes it easier to ship features. A shitty/non-existant infrastructure makes adding features exponentially harder as the system gets more complicated. Therefore, infrastructure improvements are really just features as well - as they do provide end-user value, eventually.

For example, most people would consider “User Signup” to be a feature, as a user interacts with it directly. Though, at the same time, it’s more part of your infrastructure, as it’s not the product itself, but a necessary evil to use the product. But, you’d never make signup be a manual process, as that would mean no one would sign up and you’d have no users.

Same should go for deployments - if it’s not automatic, you’ll be afraid of deploying, and favor large deployments, which will create more bugs, and a slew of other anti-patterns will become part of your process. To build a great product, you need to have an equally great infrastructure. Your users will feel the pain when your infrastructure doesn’t support your ability to enhance your product.

Think of

infrastructure

as a

product

you provide to

yourself

Whoa, far out man…

So, what would we want this infrastructure-as-a-product to do for us?

Runs the Twelve-Factor App

This methology of building web applications and services leads to more maintainable and scalable products. We’ll want our infrastructure to run these apps, as these are the apps we’ll write. Read more about 12-factor apps.
Deploy a new app

New apps should be able to be deployed to any environment within minutes. Given that we prefer micro-services over more monolithic architectures, this is an important feature - new server provisioning will happen very often. Infrastructures which requires significant effort to provision new servers tend to avoid this patter, leading to an application which is harder to scale and evolve.
Automated scaling

Every web app or service should be behind its own load-balancer, which you can easily add or remove individual servers to. The creation of these servers should be fully-automated to allow the load-balancer to create more instance for you.
Deploying should not introduce downtime

Deploying an existing app or service shouldn’t make itself unavailable during deployment. It should also allow for rollback, in case a new deployment isn’t functioning properly, and versioning, to run multiple versions of the same app.
Create and clone environments

Environments, a collection of all servers and configuration, should be easily created or cloned. This environment can be anywhere - in the cloud, or locally to develop against.
Everything is automated

Minimal manual processes should be involving in doing any of the above features. Any modifications made to a default environment should be repeatable.

Sounds awesome, right? Does it already exist? Does it kind of already exist? Let’s see what tools and services already exist that could help us out:

Heroku is “the” platform-as-a-service. It popularized many of the 12-factor app practices. It was the first to use git push to trigger deployments. Most importantly, it provides developers with a clean abstraction of an app, hiding the complexities how to run the app. While Heroku’s featureset is almost exactly what we want, this abstraction comes at a price - quite literally. Heroku is fairly expensive ($34.50/month for the low-end 2-server setup). It becomes more expensive as you use add-ons (for example, $50/month for low-end Postgres). Also, using Heroku is much more difficult (and expensive) to use for a utility server, like Jenkins or Graphite, which may have higher CPU requirements (nothing worse than slow builds).

That’s not to say if you’re using Heroku you shouldn’t, it’s a great service. Just be aware of the pros and cons.

Though, how Heroku deploys applications is very interesting. Heroku has a concept of a buildpack, which contains the logic to detect a specific type of application and deploy it. Heroku Buildpacks are all open-source, meaning you can write your own, allowing Heroku to run almost anything. We can also use Heroku Buildpacks to provision our own servers…

Docker provides application containers, which provides the applications with an isolated view of the operating system - their own process ID space, filesystem, and network interfaces. It also provides contraints on system resources (memory, CPU, network I/O).

Essentially Docker can be used to provision and deploy applications, as an alternative to Heroku Buildpacks. The difference is Docker can be used to run an identical environment locally as well as when deployed, and can run other apps like databases.

Amazon Web Services (AWS) is a massive collection of services. 37 different services to be exact:

At the core is Elastic Compute Cloud (EC2), which allows you to run and manage arbitrary virtual machines in Amazon’s data centers. EC2 also has sub-services to help route network traffic to your VMs, specifically a TCP and HTTP load-balancer, as well as Auto Scaling, which helps you launch instances and add them to a load-balancer.

Another useful service is Simple Storage Service (S3), which stores arbitrary data like a file system. It can also serve those files over HTTP, making it an ideal static HTTP server, as you don’t have to manage servers or load-balancers.

The 35 other services do lots of useful things, like various data storage services, private networks, DNS, app and mobile services, deployment/provisioning stuff, oh my! It can get very confusing for newcomers and experienced engineer alike, as some of the services overlap in functionality. For example, OpsWorks, CloudFormation, and Beanstalk all can do very similar things. Or complete with existing Open Source offerings, like DynamoDB vs. MongoDB.

Elastic Beanstalk is worth noting, because it provides a similar feature-set to Heroku - it can deploy applications to EC2 instances, and auto-configures auto scaling and load balancing. However, the pre-configured environments are awkward to configure, using the special .elasticbeanstalk and .ebextensions directories, and creating a custom AMI for Elastic Beanstalk somewhat violates the “infrastructure as code” tenant. Also, it’s all or nothing - so if your application isn’t supported by Elastic Beanstalk, like GoLang, you can’t use it.

Recently, Amazon improved Elastic Beanstalk by adding support for Docker, which can be used to avoid all the environment weirdnesses, standardize provisioning, and support other languages. It also supports rolling updates, meaning there’s no downtime during a deployment.

Amazon has recently added some new services, currently in preview, which are promising:

CodeDeploy manages applications and how to deploy them (some overlap with Beanstalk, but a bit more general)
EC2 Container Service supports clustering of Docker containers
Lambda - respond to events with code.

For those completely overwelmed by AWS, or just looking for something simpler, Digital Ocean is basically the equivalent of EC2 and Route53, but with an awesome user interface for both the website and the API, making it much easier to work with.

Because it’s only for running droplets (their word for VMs), any load balancing needs to be done with another dedicated droplet (eg. HAProxy), and scaled manually.

Feels like something's missing...

Seems like we can use all this for the major features of our infrastructure, but nothing integrates it all together and makes it as easy as Heroku.

Riker

Heroku-like application deployments for
Amazon Web Services.

Riker gives you Heroku-like application deployments for Amazon Web Services. It automates all the best-practices of application deployment, so you don’t have to piece things together.

Riker deploys any application with a buildpack to AWS, and automatically configures load balancing and auto scaling for you, as well as manages rolling deployments. It also can deploy static websites to S3.

Next I’d like to make the target “cloud” be extensible (most likely Digital Ocean as a first), support Dockerfiles as an alternative to buildpacks, and use some of AWS’s new features, like the EC2 Container Service, CodeDeploy, and Lambda.

Check out my initial post about Riker, as it’s the same content I spoke about in my talk. Here’s some other useful links:

Project on GitHub
Riker on Python Package Index
Issues list (for roadmap as well as report bugs)
Pull requests would be much appreciated!

If you want to (eventually) spend less time managing your infrastructure, and more time building products, or are just interested in making infrastructure management better, give Riker a try and come help me build it!

You're the captain, make Riker your #1

Yeah, I’m a Trekky…

Riker - Heroku-like app deployments for AWS

Wed, 24 Sep 2014 00:00:00 +0000

Riker turns Amazon Web Services into your very own Platform as a Service, giving you the same ease-of-use that platforms like Heroku provide, while retaining the control that is attractive about AWS.

Getting Started

Getting started with Riker is easy: just install it using pip and Riker will prompt you for the necessary information it needs:

$ pip install riker
$ riker config

Deploy a Node.js app

First, let’s clone a sample Node.js app:

$ git clone git@github.com:heroku/node-js-sample.git
$ cd node-js-sample

Then deploy it with just one command:

$ riker deploy

And open it in your default web browser:

$ riker open

That’s it! While this example deploys a sample Node.js application, any language can be deployed, since Riker uses Heroku Buildpacks. More information on exactly what buildpacks Riker supports, and how to use a custom buildpack.

Deploy a static website

Riker can also be used to deploy static websites to S3, rather than running nginx or apache in an EC2 instance to serve static files:

First, generate a static website:

$ mkdir static-website && cd static-website
$ echo "Hello, World" > index.html
$ touch .s3 # tells Riker to deploy to Amazon S3
$ git init && git add -A && git commit -m "Initial commit"

Then use the same commands as before to deploy and open the website:

$ riker deploy
$ riker open

Scalable, zero-downtime deployments

Riker provides a scalable mode (--scale), where load balancing and auto-scaling are configured for you. This also ensures zero-downtime during deployments.

$ riker deploy --scale

How It Works

Riker provides a set of best-practices and structure to your infrastructure, while leaning on as much of the features AWS provides as possible.

Load Balancing: Elastic Load Balancer (ELB)
Scaling: EC2 Auto-Scaling and Launch Configurations
DNS: Route 53
Static website hosting: Simple Storage Service (S3)

Riker doesn’t require its own server to be running; it’s just a deployment tool (though it’s best to run Riker in a CI server like Jenkins).

Riker also abstracts away the EC2 instances, and instead lets you only care about the application itself: code, dependencies, and configuration. Riker makes your instances easy to destroy and recreate, removing the need to ever manually provision a machine. This also makes scaling trivial. To enable this abstraction, Riker uses Docker to run your application in its own sandbox, as well as Heroku Buildpacks to deploy many different types of applications. Today, Riker also uses Dokku, but may move away from it to allow Riker to also run your application locally during development, but more about that later.

By default, Riker will deploy changes directly to existing instances. However, with the --scale flag, Riker will deploy changes to new instances, and only swap old instances out for new instances when the new instances become healthy, and the old instances no longer have active connections.

Disclaimer

While Riker is in active development, it’s probably not ready for prime-time, unless you’re really willing to help contribute and test things out.

Contribute

There are many ways you can help make Riker better:

Use it! Report bugs, suggest features and ask questions on GitHub
Write documentation: Anything you feel requires documentation, please feel free to write it in the Wiki. Please follow these instructions for contributing to Wikis
Write code: Either from the Issues list, or your own idea. Pull requests are encouraged!

Let me know what you think! Look for subsequent posts about the rationale behind Riker and why it’s different than other comparable services.

You’re the captain - make Riker your "Number One"

Splitting up a Git repo

Thu, 17 Jul 2014 18:19:00 +0000

When it comes to organizing Git repositories, there’s only one way to do it: a repository per project or component (a web server, an iOS app, etc…).

If you’ve done anything other than that, you’ve fucked up. Maybe you have a single repository for everything? Some combination? Sure, it might be easier to just have one repository, but as your system grows, the ramifications of that fuck-up become clear:

A 1-line change to the app performs a full-redeploy of your web-server - now you have to have your CI server only build certain jobs based on directory’s touched. Yuck.
Every single server has the code for all the other servers. Now you have to perform a build step to remove files from deployments. Yuck.
Large commits cutting across multiple components. Commit messages could be considered novels. Now we have to selectively commit certain files that touch a server, and certain files that touch the app, etc. Yuck.

If this sounds like your project, don’t keep working around the problem - let’s fix it. It’s probably going to be a bit of work, but it’s necessary, so take a deep breath, and let’s get started.

The first choice you have to make is whether to preserve history or not. For almost all cases you should preserve history, but if for some reason you decide it’s not important, then “these aren’t the droids you’re looking for”. You just need to create new repositories, copy files to each as appropriate, and you’re done. That might be the right choice for your project, but this isn’t the post for you.

Git gives us many tools to rewrite history, so let’s look at situations you’ll run into, and how to use Git to deal with them.

Extract a sub-directory

This is the best-case scenario: everything is perfect about your repository, only you have each project/component in a sub-directory of your large repo. For this, git subtree split is exactly what you want.

git subtree split -prefix=subdir_to_remove -b new_project_branch

For some older versions of git, the subtree subcommand may be unavailable. Instead, you can use the filter-branch subcommand:
git filter-branch --subdirectory-filter subdir_to_remove new_project_branch

Then you can push each new branch to its own remote.

Arbitrary paths

What if a project has two top-level directories, or shares files between projects? Maybe a .env or Procfile? In this case you’ll have to extract history by arbitrary paths.

To do this, first we remove all non-wanted files from all git commits with the git filter-branch command. Any commits that do not touch those paths will be removed because of the --ignore-unmatch flag.

PATHS_TO_KEEP="./some_dir ./some_other_dir .env Procfile"
git filter-branch -f --index-filter "git rm --ignore-unmatch --cached -qr -- . && git reset -q \$GIT_COMMIT -- $PATHS_TO_KEEP" --prune-empty -- HEAD

The -- HEAD means to only do this transformation in the current branch. You can replace it with -- --all to do it across all branches. Or --tag-name-filter cat -- --all to do it across all branches AND tags. Pick your poison.

While this is great, empty merge commits are not removed by the last command! A parent-filter can be used to rewrite commit parents to remove them:

tmpfile=`mktemp -t splitgit`
chmod 755 $tmpfile
echo '#!/usr/bin/env ruby
old_parents = gets.chomp.gsub("-p ", " ")
new_parents = old_parents.empty? ? [] : `git show-branch --independent #{old_parents}`.split
puts new_parents.map{|p| "-p " + p}.join(" ")' > $tmpfile
git filter-branch -f --prune-empty --parent-filter $tmpfile HEAD
rm $tmpfile

You can simplify this by writing $tmpfile to disk in a well-known location, instead of using mktemp each time.

Moving files around

Let’s say a new repo should still have a file, but you want to change its path in the new repo? If so, tree-filter can help with this. For this example, we want to move an entire sub-directory up one level.

git filter-branch -f --tree-filter 'shopt -s dotglob nullglob; test -d sub_dir && mv sub_dir/* . || :' HEAD

Deleting files

What if you just don’t want a particualr file around?

git filter-branch -f --index-filter 'git rm --cached --ignore-unmatch Rakefile' --prune-empty -- HEAD

We use index-filter here to not require files be checked out for each commit, which makes it faster. This is different from tree-filter in that it checks out each commit’s files. This is good in that it lets you use rm and test, but is slower. When formulating your own history manipulations, keep this in mind.

Keeping only certain lines in a file

(I know, this is crazy…) What if, in your old repo, you had a configuration file that was for all projects, but now you want to only keep commits that affected the variables for the particuar sub-project? tree-filter + cat + grep:

git filter-branch -f --tree-filter 'test -f .env && cat .env | grep "\(^VAR_X\|^VAR_Y\|^VAR_Z\)" > .newenv && mv .newenv .env || :' --prune-empty -- HEAD

Fucking symlinks

OK, this is harder, but all your fault. You have symlinks in your repo, presumably to share code between projects. Bad. Share code through a package manager, a submodule, or some other means, but not through a symlink.

This is a problem when your repository split makes a symlink invalid, since Git stores symlinks as a file with its only contents a path.

To fix this, we’ll convert all symlinks into hard links, which Git will then treat as any other file and commit its contents.

Convert all symlinks to hard links (if the symlinks are relative paths):

git filter-branch -f --tree-filter \
'find . -type l -exec bash -c '"'"'ln -f "$(dirname "$0")/""$(readlink -n "$0")" "$0"'"'" {} \;' \
--prune-empty -- HEAD

If the paths are absolute, WTF? How the hell did that ever work to begin with??

Empty commits

If for some reason you have empty commits (maybe you forgot a --ignore-unmatch ?), you can remove them with a special --commit-filter:

git filter-branch --commit-filter 'git_commit_non_empty_tree "$@"' HEAD

The initial commit is still empty, what gives?

That’s the only price your going to pay for not doing the right thing to start, so be OK with it. Your initial commit is most likely going to be an empty one saying “Initial commit” or whatever you said.

You could rebase your entire repository (git rebase --interactive --root) but if you have any legitimate merge commits you’ll find yourself attempting to merge years old code in the right order. It’s not worth it.

Fixing author info

If you notice certain committers used a non-company email, or there are commits from non-human accounts, you can use a --env-filter to rewrite author and committer name and emails. GitHub’s Help site has a good example of this.

Holy crap I deleted the wrong thing/everything!

No problem! If anything goes wrong, you can undo the last filter-branch:

git reset --hard refs/original/refs/heads/master

Next steps

After you’ve got all your new shiny respositories how you want them, you’re may not be entirely done. Some things that you probably need to do:

Create new CI or deployment jobs to work against new repos
Reference GitHub issues in commits? Well, now #1234 points to your new repo, not the old one. To reference an issue in a different repo than the current one, use user/repo#1234 (and make it a habit to always reference them like that). I’ll leave it to the reader to figure out how to rewrite commit messages from #1234 to user/repo#1234.
Delete old code from old repo, and put a message up on your old repo telling visitors where to look for the new repo.

Happy history rewriting!

Sparklines in WPF and Silverlight

Mon, 09 May 2011 04:15:00 +0000

One seemingly-trivial-yet-recurring problem in financial software is the need for a live-updating line-chart. However, from multiple conversations with Lab49 folks, as well as from experience during my first project, I’ve learned that most WPF/Silverlight charting packages fall short in various ways, especially if you’re updating their data frequently. Seems like everyone just rolls their own line chart and tailors it to each project, but doesn’t share it for some reason. I’d like to break that trend by sharing and early version of my own sparkline control for WPF and Silverlight.

http://github.com/jschementi/sparkline

Sample Usage:

It’s implementation is very basic; Sparkline.AddTimeValue constructs a point at the next time interval and adds it to a Polyline. You can control the sparkline’s visuals, including adding visible points along the line and showing horizontal lines for the latest/highest/lowest values. The source builds assemblies for both .NET 4.0 and Silverlight 4.

There are obvious features missing like rendered axis or variable x-axis (time) values, but hopefully this provides a simple starting place for anyone else needing a very simple updating line graph.

By the way, Andre de Cavaignac, a colleague of mine at Lab49, and Daniel Simon shared their own a while back: Live Updating Line Graph in WPF. Let me know if there are any others out there.

MacDrive 7 with Windows 7

Tue, 12 May 2009 21:23:00 +0000

UPDATE: MacDrive 8 has been released, which has support for Windows 7, so the work-around explained in this post is no longer necessary. Also, people have been experiencing issues with it, so my advice is to upgrade.

If you’re running Windows on a Mac, and you’re shuffling files between Mac OS and Windows, MacDrive is the best tool for mounting your HFS (Mac OS’s default file-system format) partitions in Windows. MacDrive costs $49.95, but is well worth it. However, MacDrive does not release versions for pre-release operating systems, and there are a couple speed bumps when trying to use MacDrive on Windows 7. There are various version of this solution posted in a bunch of forums, but no definitive guide. Here’s exactly how I solved it:

MacDrive’s installer will only run on Windows XP and Vista.
To have it run on Windows 7, you must run as Administrator, run in “Windows Vista” compatibility mode.
Also remove the operating system version checks the installer makes with Orca. MacDrive’s installer is an self-extracting exe, so you must extract it first; using 7-zip or any other unzipping tool will do the trick. Once extracted launch Orca and open MacDrive/x86.en-US/MacDrive7.x86.en-US.msi (or the x64 version if that applies to you). Once open, look for the “LaunchCondition” table and delete all rows from it.
If you run the main setup.exe from the extracted location it will install fine, but upon reboot you will not see your Mac OS drive. This is because MacDrive simply doesn’t assign the drive a letter. To do so, download Ext2Fsd, unzip, and run Ext2Mgr.exe. Find your HFSJ-formatted partition, right-click and select “Change Drive Letter”. Assign it a letter, exit the application.
Open up your “Computer” screen and you should see your Mac OS drive now:

Please post in the comments if this doesn’t work for you, and I’ll update the steps. Enjoy!

Git on Windows 7

Mon, 12 Jan 2009 03:52:00 +0000

I use Git everyday, whether it be for IronRuby, AgDLR, or random other Rails/Ruby work. Being that I work at Microsoft, I got my hands on earlier build of Windows 7 during the fall. Unfortuantely, git didn’t work fully on Windows 7, failing whenever a merge happened, which made push fail, as well as pull when trying to get other people’s changes). Jim Deville has been tracking this with the Windows team, and even bringing it up to the msysgit maintainer to see if there’s any way around this. It was looking grim, so I left my main development machine on Vista, and was sad.

Windows 7 Beta was released this past Friday, and I managed to download the .iso, get a Product Key, etc, at home during the day on Saturday. Now that it’s released, I’m guessing other people will have this problem ...

I’m not sure of the specific details behind the issue, but msysgit calls into C:\windows\system32\msvcrt.dll during a merge, and the changes to that DLL in Windows 7 breaks msysgit. Last week, John Lam was describing the issue to John Messerly, and realized that the easiest fix would be to copy the Windows Vista msvcrt.dll to the C:\Program Files\Git\bin directory, since Windows DLL resolution will look on the PATH, and “.” is included.

Curt Hagenlocher just reminded me of the fix, and it works!

Pretty awesome. Enjoy Git in Windows 7!

Live Mesh doesn't require UAC!

Wed, 18 Jun 2008 21:25:00 +0000

Today my three machines (2 Vista and 1 XP) which use Live Mesh asked to be updated. No information to what the updates include ... just that they wanted to plug back into the mother ship. I thought, "Could this be, the infamous Live Mesh update that removes the need for UAC"? After updating my Vista machines, I disabled UAC, rebooted, and ... Live Mesh still worked fine! Woohoo!

This should now convince more people to install and use Live Mesh.

PS: I searched the tubes for any announcement of this awesome update, and found nothing. Nothing. This guy had the same experience. However, today (a day after the update) at 11:30a a blog entry appeared on the Live Mesh blog announcing the service update.

Now, I suffer from the same delay-between-release-and-blogging syndrome that most project owners suffer from, but all projects are expected to have some explanation of what the release contains ... like release notes, etc. Live Mesh uses there blog as the release notes, so they have more of a need to be on-top of their blog.

Fixing Mac OS X - Leopard's Spaces

Mon, 31 Mar 2008 02:34:00 +0000

If anyone has turned on "Spaces" in Leopard, you'll know why Apple decided to ship it turned off; the window switcher (Command-Tab) doesn't understand Spaces, and will automatically warp you to the first space that has that application running on it. In practice, this is horrible if you have apps running on different spaces (browsers, terminal, textmate, etc). Luckily, Apple has a way of changing that behavior! In 10.5.2 they introduced a default option to disable space switching on Command-Tab. Open up Terminal or iTerm and type the following:

defaults write com.apple.Dock workspaces-auto-swoosh -bool NO
killall Doc

This will make Command-Tab only switch between applications running in the same Space. However, if you switch to an Application which is not in your current Space, it will ~~still warp you to the Space where it is~~ not warp you between spaces; you much always switch spaces manually. Regardless, it's much better and makes Spaces much more usable. Though, Workspaces in Gnome still feel much better.

NTFS Symbolic Link

Sun, 23 Dec 2007 06:51:00 +0000

I almost hate to say it, but the more I use Vista for my day-to-day work the more I hate it less. Yes, that was a lot of small hates.

Vista now supports symbolic/hard links like POSIX systems do (*nix, bsd, etc). This is amazing, as I use them like crazy on my Ubuntu and Leopard machines.

It's as easy as:

C:\>mklink /d bar c:\windows

More info: http://en.wikipedia.org/wiki/NTFS_symbolic_link

I stumbled upon this while trying to set up the Zune software and iTunes (on Leopard) to pull from the same music library ... on yet another machine. On Leopard, I simply put a symlink to where my music library is mapped, so it looked something like this:

$/Users/jimmy>ln -s /Volumes/jimmy/Music/iTunes Music/iTunes

This requires I have my shared library mounted to that volume already, so I created a little automator script to mount smb://jimmy-zeus/jimmy to /Volumes/jimmy.

In Vista, this is simply one step. You can make the symlink directly to a SMB file-share! (Please someone correct me if you can do this in *nix ... I'll promise to feel like an idiot =P).

C:\Users\jimmysch>mklink /D iTunes \\jimmy-zeus\jimmy\Music\iTunes\iTunes Music

Pretty freakin' cool!

[Update] Ok, maybe SMB file-share symlinks aren't so cool; as explorer.exe doesn't understand them and will hang the UI if the server doesn't respond right away when navigating to a folder with ANY SMB SYMLINK'd FOLDERS! And if you don't have access to the server ... you must wait for the 60 second timeout to get control of the window again. Ugh. Use wisely ...

Upgrading Assistment to Rails 2.0

Sun, 16 Dec 2007 08:56:00 +0000

image linked from b.lesseverything.com

Today I spent some time upgraded Assistment (http://assistment.org) to Rails 2.0. I made the mistake of upgrading Rails before changing anything, and then fixing the errors one by one ... oy. So, I'll quickly go over what I had to do to get stuff working again.

Fix deprecation warnings in 1.2.6

If you haven't upgraded yet, make sure you're on Rails 1.2.6 and get rid of your deprecated API calls: http://www.rubyonrails.org/deprecation. To easily find your deprecated usages grab the rake task from this site: http://www.slashdotdash.net/articles/2007/12/03/rails-2-upgrade-notes. And of course, make sure your tests pass and your app actually runs before moving on!

Keep in mind some pretty useful parts of Rails have been pulled out into plugins, which you can find at http://svn.rubyonrails.org/rails/plugins/. I had to install acts_as_list, acts_as_tree, classic_pagination (though, we really need to move to will_paginate!), and in_place_editing. I also had to spend time updating my forms to use the block-notation rather that start/end form tags. That was annoying. One plugin I had trouble with was acts_as_paranoid. It calls construct_count_options_from_legacy_args, which is deprecated, but I have not found an update to the plugin. To fix it, I had to add that method to the plugin. I left the deprecation notice in there to remind me to fix it =) Also, if you used strip_tags in your model (shush, I had my reason: I used it since I stored HTML in the database and I wanted a nice way to clean it), you have to include ActionView::Helpers::SanitizeHelper in your model; just including ActionView::Helpers::TextHelper won't cut it.

Upgrading Rails

gem update rails

should do the trick for upgrading you. I had some problems upgrading on Leopard, so I had to gem remove rails and the active*/action* gems before running "gem install rails". Even that failed the first 2 times with 404 errors for finding certain gems, but the 3rd time it worked.

rake rails:update

Run that in your project to upgrade your configs, javascripts, and scripts. I took the extra step and generated a new rails project and merged any new config changes into Assistment. This was mainly environment.rb, and the extremely useful addition of a initializers directory for application config.

Running Two point Oh!

(Note: to use ruby-debug you don't need to manually set it up in environment.rb. Simply run ./script/server --debugger to enable debugging with the "debugger" method. Yep, that simple.) Run your app/tests and your on Rails 2.0! If anyone has any additional issues, google is your friend.

Rounded corners with Transcorners

Tue, 24 Apr 2007 23:42:00 +0000

I just spent the past hour and a half trying to get simple rounded corners using Transcorners, and I wanted to let everyone know about a little "gotcha" so no one else will waste all that time like I did. If you are experiencing the background of your rounded div not being rounded ... and the border BOTH rounding and not ... ... then you are experiencing the same problem I did. To fix this, simply make sure the background-color of one of the parent elements is set. In my case, nothing had it set, including the body ...

body {
 background-color: white;
}

... this will make sure the background color is set and you will achieve the correct rounded corner effect:

Book Reviews - Head First OOA&D and Rails Recipies

Thu, 22 Feb 2007 18:48:00 +0000

Check out my review of Head First Object-Oriented Analysis & Design on IBMs Rational Edge e-zine.

Head First OOA&D Review

Also, I don't think I ever posted about my review of Rails Recipes, check that out!

Rails Recipes Review

Assistments

Mon, 19 Feb 2007 02:51:00 +0000

This year, I'll be working as the Lead Developer in Neil Heffernan's Artificial Intelligence Lab, which is part of WPI's Computer Science Department. This lab isn't a traditional lab, they actually make a product! It's called ASSISTments, an intelligent tutoring system that's being used in MA schools today. Undergraduate and graduate students from WPI primarily built it, but Carnegie Mellon also collaborates in research.

We're working on a new version of the system to enable more scaffolding of questions, more question types, and just faster development in general. The current system is written in Java, and we're rebuilding it with Ruby on Rails. Some of the new features we're adding are effort calculations, content feedback, calculator support, content versioning. Check out the progress blog at http://assistment.blogspot.com.

This effort will be a case study for my master's thesis about productizing a research project.

Nichum.com

Sun, 24 Dec 2006 01:34:00 +0000

I'm launching a website with Alex Kantrowitz that will help a lot of people in mourning.

Check out Nichum.com to see what it's about.

Articles for CLR Inside Out

Wed, 06 Sep 2006 05:26:00 +0000

I got two articles published in MSDN Magazine!

Starline

Wed, 19 Jul 2006 09:10:00 +0000

Yesterday I launched Starline (starline.schementec.com), a web-based point-of-sale app. It was initially written in Visual Basic for a specific restaurant, a project Mike Veilleux started and I continued. We’ve moved between so many different iterations, but now it's landed on a web app, and feels much better. Check it out!

Hard Drive Recovered!

Fri, 07 Oct 2005 15:25:00 +0000

Follow up from previous post

I was about to go to Fedex to send my hard drive, but I just couldn’t believe that this was my only option. To cut to the chase, I got all my data back!!! All thanks to Active@ UNDELETE!

Active@ is a Software Company specializing in data recovery. They make 2 tools that I used, Partition Recovery and UNDELETE, to save my files. I really only needed UNDELETE but Partition Recovery let me see my files and give me hope that a bomb hitting my room is much worse.

I also bought a Maxtor OneTouch II 300GB portable hard drive. This thing is pretty sweet. It not only a lot of storage, but it comes with Retrospect which will let you schedule backups. It also has this button on the front of the drive that when pressed will backup and newly created/changed files. Sweet stuff.

Anyway, now I can get on with my life. Back up regularly. definitely a huge learning experience for me about how hard drives work.

Hard Drive Failure

Wed, 05 Oct 2005 19:04:00 +0000

Friday night my computer started to act weird. I backed up all my data to another location of the drive where it would be safe, but I didn’t think to get it off my computer and onto something else like another hard drive or DVDs. That was my mistake. Something went wrong and I lost access to all my files. There are still on my hard drive but I can’t get to them.

I’m thinking of sending the hard drive to a data recovery company to see if they can fix it, but it’s going to cost more than $1K to recover the data.

Moral of the story … back up your data!

Update Got my data restored!

Ode to Joystick - Newspaper Interview

Tue, 23 Nov 2004 03:21:00 +0000

Ode to joystick - Worcester Telegram and Gazette.

Here’s a mirror of the article about the new Interactive Media and Game Development (IMGD) major at WPI. Being the head of the Game Development Club at WPI, Chris and I were interviewed for our take on the major.

Jimmy Schementi

Get Your Infrastructure Right

Have you ever...

Changed code

directly in

production?

Application (feature)

vs.

Infrastructure (chore)

Prioritizing

infrastructure

makes it easier to

ship features

Think of

infrastructure

as a

product

you provide to

yourself

Feels like something's missing...

Riker

Heroku-like application deployments forAmazon Web Services.

You're the captain, make Riker your #1

Riker - Heroku-like app deployments for AWS

Getting Started

Deploy a Node.js app

Deploy a static website

Scalable, zero-downtime deployments

How It Works

Disclaimer

Contribute

Splitting up a Git repo

Extract a sub-directory

Arbitrary paths

Moving files around

Deleting files

Keeping only certain lines in a file

Fucking symlinks

Empty commits

The initial commit is still empty, what gives?

Fixing author info

Holy crap I deleted the wrong thing/everything!

Next steps

Sparklines in WPF and Silverlight

MacDrive 7 with Windows 7

Git on Windows 7

Live Mesh doesn't require UAC!

Fixing Mac OS X - Leopard's Spaces

NTFS Symbolic Link

Upgrading Assistment to Rails 2.0

Rounded corners with Transcorners

Book Reviews - Head First OOA&D and Rails Recipies

Assistments

Nichum.com

Articles for CLR Inside Out

Starline

Hard Drive Recovered!

Hard Drive Failure

Ode to Joystick - Newspaper Interview

Heroku-like application deployments for
Amazon Web Services.