Tag Archives: Riak

Using Innostore with Riak

February 22, 2010

Innostore is an Erlang application that provides an API for storing and retrieving key/value data using the InnoDB storage system. This storage system is the same one used by MySQL for reliable, transactional data storage. It’s a proven, fast system and perfect for use with Riak if you have a large amount of data to store. Let’s take a look at how you can use Innostore as a backend for Riak.

(Note: I assume that you have successfully built an instance of Riak for your platform. If you built Riak from source in ~/riak, then set $RIAK to ~/riak/rel/riak.”)

We first get started by grabbing a stable release of Innostore. You’ll need to download the source for a release from: https://github.com/basho/innostore

Looking in the “Tags & snapshots” section, you should download the source for the highest available RELEASE_* tag. In my case, RELEASE_4 is the most recent release, so I’ll grab the bz2 file associated with it.

Once I have the source code, it’s time to unpack it and build:

$ tar -xjf innostore-RELEASE_4.tar.bz2

$ cd innostore

$ make

Depending on the speed of the machine you are building on, this may take a few minutes to complete. At the end, you should see a series of unit tests run, with the output ending:

All 7 tests passed.

100222 7:43:58 InnoDB: Shutdown completed; log sequence number 90283

Cover analysis: /Users/dizzyd/src/public/innostore/.eunit/index.html

Now that we have successfully built Innostore, it’s time to install it into the Riak distribution:

$ ./rebar install target=$RIAK/lib

If you look in the $RIAK/lib directory now, you should see the innostore-4 directory alongside a bunch of .ez files and other directories which compose the Riak release.

Now, we need to tell Riak to use the Innostore driver as a backend. Make sure Riak is not running. Edit $RIAK/etc/app.config, setting the value for “storage_backend” as follows:

{storage_backend, innostore_riak},

In addition, append the configuration for the Innostore application after the SASL section:

{sasl, [ ....

]}, %% < -- make sure you add a comma here!!

{innostore, [

{data_home_dir, "data/innodb"}, %% Where data files go

{log_group_home_dir, "data/innodb"}, %% Where log files go

{buffer_pool_size, 2147483648} %% 2G in-memory buffer in bytes


You may need to adjust the directories for your data_home_dir and log_group_home_dirs to match where you want the inno data and log files to be stored. If possible, make sure that the data and log dirs are on separate disks — this can yield much better performance.

Once you’ve completed the changes to $RIAK/etc/app.config, you’re ready to start Riak:

$ $RIAK/bin/riak console

As it starts up, you should see messages from Inno that end with something like:

100220 16:36:58 InnoDB: highest supported file format is Barracuda.

100220 16:36:58 Embedded InnoDB started; log sequence number 45764

That’s it! You’re ready to start using Riak for storing truly massive amounts of data.


Dave Smith

Calling all Rubyists – Ripple has Arrived!

February 11, 2010

The Basho Dev. Team has been very excited about working with the Ruby community for some time. The only problem was we were heads down on so many other projects that it was hard to make any progress. But, even with all that work on our plate, we were committed to showing some love to Rubyists and their frameworks.

Enter Sean Cribbs. As Sean details in his latest blog post, Basho and the stellar team at Sonian made it possible for him to hack on Ripple, a one-of-a-kind client library and object mapper for Riak. The full feature set for Ripple can be found on Sean’s blog, but highlights include a DataMapper-like API, an easy builder-style interface to Map/Reduce, and near-term integration with Rails 3.

And, in case you need any convincing that you should consider Riak as the primary datastore for your next Rails app, check out Sean’s earlier post, “Why Riak should power your next Rails app.”

So, if you’ve read enough and want to get your hands on Ripple, go check it out on GitHub.

If you don’t have Riak downloaded and built yet, get on it.

Lastly, you are going to be seeing a lot more Riak in your Ruby. So stay tuned because we have some big plans.



The Release Riak 0.8 and JavaScript Map/Reduce

February 3, 2010

We are happy to announce the release of Riak 0.8 available for download immediately. Riak 0.8 features a number of enhancements to the core map/reduce machinery that will make Riak more accessible to a wider audience. The biggest enhancement is the ability to write map/reduce queries in JavaScript. We’re using our erlang_js project to integrate Mozilla’s Spidermonkey engine directly into Riak to keep overhead to a minimum.

We’ve also built a spiffy REST API for submitting map/reduce queries. Queries are described in JSON and POST-ed to the Riak server. Results are sent back as JSON for your processing pleasure. And, the REST interface supports streaming results for large result sets, too.

To kick it all off, we’ve put together a short screencast demonstrating how to use Riak’s flashy new features. You can watch it below, or view it on Vimeo. There’s also a slew of bug fixes and optimizations included in Riak 0.8. See the release notes for all the juicy details.

Download and enjoy!

View on Vimeo

Basho Podcast Three – An Introduction To Innostore

February 2, 2010

You may remember that Basho recently open-sourced Innostore, our standalone Erlang application that provides a simple interface to embedded InnoDB…

In this podcast, Dave “Dizzy” Smith and Justin Sheehy discuss the release of Innostore, why we built it, how we use it in Riak, and why it might be useful for other Erlang projects. The discussion focuses on the stability and predictability of InnoDB, especially under load and as compared with other storage backends like DETS.

And of course, go download Innostore when you are done with the podcast.



If you are having problems getting the podcast to play, click here to play in new window or right click to download the podcast.

Why Vector Clocks are Easy

January 29, 2010

Vector clocks are confusing the first time you’re introduced to them. It’s not clear what their benefits are, nor how it is you derive said benefits. Indeed, each Riak developer has had his own set of false starts in making them behave.

The truth, though, is that vector clocks are actually very simple, and a couple of quick rules will get you all the power you need to use them effectively.

The simple rule is: assign each of your actors an ID, then make sure you include that ID and the last vector clock you saw for a given value whenever to store a modification.

The rest of this post will explain why and how to follow that simple rule. First, I’ll explain how vector clocks work with a very simple example, and then show how to use them easily in Riak.

Vector Clocks by Example

We’ve all had this problem:

Alice, Ben, Cathy, and Dave are planning to meet next week for
dinner. The planning starts with Alice suggesting they meet on
Wednesday. Later, Dave discuss alternatives with Cathy, and they
decide on Thursday instead. Dave also exchanges email with Ben, and
they decide on Tuesday. When Alice pings everyone again to find out
whether they still agree with her Wednesday suggestion, she gets
mixed messages: Cathy claims to have settled on Thursday with Dave,
and Ben claims to have settled on Tuesday with Dave. Dave can’t be
reached, and so no one is able to determine the order in which these
communications happened, and so none of Alice, Ben, and Cathy know
whether Tuesday or Thursday is the correct choice.

The story changes, but the end result is always the same: you ask two people for the latest version of a piece of information, and they reply with two different answers, and there’s no way to tell which one is really the most recent.

Vector clocks to the rescue, but how? Simple: tag the date choice with a vector clock, and then have each party member update the clock whenever they alter the choice. Start with Alice’s initial message:

date = Wednesday
vclock = Alice:1

Alice says, “Let’s meet Wednesday,” and tags that value as the first version of the message that she has seen. Now Dave and Ben start talking. Ben suggests Tuesday:

date = Tuesday
vclock = Alice:1, Ben:1

Ben left Alice’s mark alone, but added a mark specifying that it was the first version of the message that he had seen. Dave replies, confirming Tuesday:

date = Tuesday
vclock = Alice:1, Ben:1, Dave:1

Just like Ben’s modification, Dave just adds his own first-revision mark. Now Cathy gets into the act, suggesting Thursday:

date = Thursday
vclock = Alice:1, Cathy:1

But wait, what happened to Ben’s and Dave’s marks? Cathy didn’t have a version of the object that had been modified by Ben or Dave, so their marks can’t appear in her modification. This means that Dave has two conflicting objects:

date = Tuesday
vclock = Alice:1, Ben:1, Dave:1


date = Thursday
vclock = Alice:1, Cathy:1

Dave can tell that these versions are in conflict, because neither vclock “descends” from the other. In order for vclock B to be considered a descendant of vclock A, each marker in vclock A must have a corresponding marker in B that has a revision number greater than or equal to the marker in vclock A. Markers not contained in a vclock can be considered to have revision number zero. So, since the Tuesday value has a Cathy revision of zero while Thursday has a Cathy revision of one, Tuesday cannot descend from Thursday. But, since Thursday has Ben and Dave revisions of zero while Tuesday has Bend and Dave revisions of one, Thursday is also not descended from Tuesday. Neither succeeds the other, so Dave has a conflict to sort out.

Luckily, Dave’s a reasonable guy, and chooses Thursday:

date = Thursday
vclock = Alice:1, Ben:1, Cathy:1, Dave:2

Dave also created a vector clock that is successor to all previously-seen vector clocks: it has revision numbers for every actor equal to or greater than the last revision number he saw for that actor. He emails this value back to Cathy.

So now when Alice asks Ben and Cathy for the latest decision, the replies she receive are, from Ben:

date = Tuesday
vclock = Alice:1, Ben:1, Dave:1

and from Cathy:

date = Thursday
vclock = Alice:1, Ben:1, Cathy:1, Dave:2

From this, she can tell that Dave intended his correspondence with Cathy to override the decision he made with Ben. All Alice has to do is show Ben the vector clock from Cathy’s message, and Ben will know that he has been overruled. (Dave will, almost certainly, blame his broken email software for failing to inform Ben of the change.)

How to do this in Riak

Now that you understand vector clocks, using them with Riak is easy. I’ll use the raw HTTP interface to illustrate.

First, whenever you store a value, include an X-Riak-ClientId header to identify your actor. For Alice’s first message above, you’d say:

curl -X PUT -H "X-Riak-ClientId: Alice" -H "content-type: text/plain" 
http://localhost:8098/raw/plans/dinner --data "Wednesday"

When Ben, Cathy, and Dave each GET Alice’s plans, they’ll get the same vector clock (I’ve removed some of the other headers for brevity):

curl -i http://localhost:8098/raw/plans/dinner
HTTP/1.1 200 OK
X-Riak-Vclock: a85hYGBgzGDKBVIsrLnh3BlMiYx5rAzLJpw7wpcFAA==
Content-Type: text/plain
Content-Length: 9


The X-Riak-Vclock header contains an encoded version of a vclock that is the same as out earlier example: Alice has modified this value once.

Now when Ben sends his change to Dave, he includes both the vector clock he pulled down (in the X-Riak-Vclock header), and his own X-Riak-Client-Id:

curl -X PUT -H "X-Riak-ClientId: Ben" -H "content-type: text/plain" 
-H "X-Riak-Vclock: a85hYGBgzGDKBVIsrLnh3BlMiYx5rAzLJpw7wpcFAA==" 
http://localhost:8098/raw/plans/dinner --data "Tuesday"

Dave pulls down a fresh copy, and then confirms Tuesday:

curl -i http://localhost:8098/raw/plans/dinner
X-Riak-Vclock: a85hYGBgymDKBVIsrLnh3BlMiYx5rAymfeeO8EGFWRLl30GF/00ACmcBAA==
curl -X PUT -H "X-Riak-ClientId: Dave" -H "content-type: text/plain" 
-H "X-Riak-Vclock: a85hYGBgymDKBVIsrLnh3BlMiYx5rAymfeeO8EGFWRLl30GF/00ACmcBAA==" 
http://localhost:8098/raw/plans/dinner --data "Tuesday"

Cathy, on the other hand, hasn’t pulled down a new version, and instead merely updated the plans with her suggestion of Thursday:

curl -X PUT -H "X-Riak-ClientId: Cathy" -H "content-type: text/plain" 
-H "X-Riak-Vclock: a85hYGBgzGDKBVIsrLnh3BlMiYx5rAzLJpw7wpcFAA==" 
http://localhost:8098/raw/plans/dinner --data "Thursday"

(That’s the same vector clock that Ben used, in that encoded gibberish is making your eyes cross.)

Now, when Dave goes to grab this new copy (after Cathy tells him she has posted it), he’ll see one of two things. If the “plans” Riak bucket has the allow_mult property set to false, he’ll see just Cathy’s update. If allow_mult is true for the “plans” bucket, he’ll see both his last update and Cathy’s. I’m going to show the allow_mult=true version below, because I think it illustrates the flow better.

curl -i -H "Accept: multipart/mixed" http://localhost:8098/raw/plans/dinner
HTTP/1.1 300 Multiple Choices
X-Riak-Vclock: a85hYGBgzWDKBVIsrLnh3BlMiYx5rAymfeeO8EGFWRLl30GF1fsRwsypF59BhT0mIoTZ/1SYQIUrEcJszUksu9R6kCWyAA==
Content-Type: multipart/mixed; boundary=ZZ3eyjUllBi7GXRRMJsUublFxjn
Content-Length: 368

Content-Type: text/plain

Content-Type: text/plain


Dave sees two values because the vclock that Cathy generated wasn’t a successor to the vclock that Dave had generated with his last modification. Riak couldn’t choose between them, and therefore kept both values.

Dave picks Thursday, and updates the object, resolving the conflict. Riak has already computed a unified, descendant vector clock for Dave, so he uses the vector clock from the multi-value version he just pulled down, just like before:

curl -X PUT -H "X-Riak-ClientId: Dave" -H "content-type: text/plain" 
-H "X-Riak-Vclock: a85hYGBgzWDKBVIsrLnh3BlMiYx5rAymfeeO8EGFWRLl30GF1fsRwsypF59BhT0mIoTZ/1SYQIUrEcJszUksu9R6kCWyAA==" 
http://localhost:8098/raw/plans/dinner --data "Thursday"

Now when Alice check for the latest version, she just sees the final decision:

curl -i http://localhost:8098/raw/plans/dinner
HTTP/1.1 200 OK
X-Riak-Vclock: a85hYGBgzWDKBVIsrLnh3BlMiYx5rAymfeeO8EGFWRLl30GF1fvhwmzNSSy71HqgEpUTEerZ/1SYYBFmTr34DCjMBBTOnQwUzgIA
Content-Type: text/plain
Content-Length: 7


While Riak couldn’t decide whether to choose Cathy’s modification over Dave’s earlier modification, it was easy to choose Dave’s latest modification, because the vclock created was a successor to the vclock in place.


So, vclocks are easy: assign each of your actors an ID (“Alice”, “Ben”, “Cathy”, and “Dave” in these examples), then make sure you include that ID and the last vector clock you saw for a given value whenever to store a modification.

If two actors store changes with vector clocks that don’t descend from each other, Riak will store and hand back both values. When descendancy can be calculated, values stored with vector clocks that have been succeeded will be removed.


Innostore — connecting Erlang to InnoDB

January 26, 2010

Riak has pluggable storage engines, and so we’re always on the lookout for better ways that users can store their data locally. Recent experiences with some Basho customers managing some large datasets led us to believe that InnoDB might work out very well for them.

To answer that question and fill that need, Innostore was written. It is a standalone Erlang application that provides a simple interface to Embedded InnoDB. So far its performance has been quite good, though InnoDB (with or without the Innostore API) is highly dependent on tuning the local configuration to match the local hardware. Luckily, Dizzy – the author of Innostore — has some heavy-duty experience doing that kind of tuning and as a result we’ve been able to help people meet their performance goals using Innostore.


Basho Podcast Two – An Introduction to erlang_js

January 19, 2010

Check out the new Basho podcast featuring Kevin Smith and Bryan Fink that discusses erlang_js, a simple and easy-to-use binding between Erlang and JavaScript. It is packaged as an OTP application so developers can easily embed Javascript inside their own applications.

Once you are done with the podcast, go download erlang_js.


Mark Phillips

Right click here to download the Podcast

A Quick Note on Rebar

January 10, 2010

As many of you Erlang and Riak fans know, Dave Smith has been hard at work on Rebar. For those of you who don’t know, Rebar is a truly cool packaging and build tool for Erlang applications. Dave took a break from coding this morning to post a few words on his blog Gradual Epiphany about why he was inspired to write Rebar and what it means for building and deploying applications. Check it out. It’s a great read.

Also, if you haven’t had a chance to join the Rebar mailing list, you can do so here.


Riak Screencasts and Presentations – The Collected Works

December 28, 2009

To date, there have been a number of screencasts and presentations done on Riak and Riak-related technologies. As a belated holiday gift (we were coding, not blogging), we thought it would be a valuable resource if we assembled all of them in an easy-to-peruse list here on the Basho Blog. If we missed any, please let us know in the comments.

Go forth and consume!