Sharing is caring: Ruby, Perl, Memcached and MsgPack

Do you need to share data structures between ruby and perl ... FAST?

I recently saw MsgPack bubble through my RSS feeds and tagged it to go have another look. It provides very fast multi-language bindings for serialisation/de-serialisation.

A few quick experiments with the data I share (via memcached) between ruby and perl showed a write (serialisation + write to memcached) speed increase from 20s to 1.8s. Read (read from memcached + de-serialisation) performance showed similar performance increases.

All initial testing was done ruby -> memcached -> ruby but as soon as I switched to reading from memcached via perl I started getting 'extra bytes' errors from the perl side. I then tried perl -> memcached -> perl and everything was fine.

Weird.

A closer look at the data written to memcached and then read from perl showed that the data serialised with MsgPack on the ruby end was not the same as the data read by perl from memcached (validating the 'extra bytes' error).

Testing the write -> read process from perl to ruby yielded the following error:

/opt/local/lib/ruby/gems/1.8/gems/memcached-0.19.5/lib/memcached/memcached.rb:514:in `load': incompatible marshal file format (can't be read) (TypeError)

format version 4.8 required; 147.1 given

from /opt/local/lib/ruby/gems/1.8/gems/memcached-0.19.5/lib/memcached/memcached.rb:514:in `get'

from ./t_msgpack.rb:35:in `read_test'

from ./t_msgpack.rb:49

Now why on earth would I be getting a 'incompatible marshal file format' error as I am not using the ruby marshalling lib at all?

Turns out the memcached lib I use turns marshalling of ruby data on by default when you write to/read from memcached. This is most likely the best option for most cases where you don't want to use some other form of serialisation/de-serialisation but was really biting me here.

The solution is to simply stop the default behaviour of the memcached lib by using the following forms of get and set that turns on the 'raw' data handling switch for the memcached lib:

get KEY, false

set KEY, VALUE, TTL, false

The 'false' parameter at the end of those overrides the default behaviour turning default serialisation/de-serialisation via Marshall off.

Reality restored.

Broken memcached gem/libmemchached deb on Ubuntu

This post follows on a previous post that was OS X centric.

To fix this issue follow the same steps (starting with the download of libmemcached-0.25.14.tar.gz) detailed here.

Because the source installation of libmemcached dropped the libs in a non-standard place (from a Debian/Ubuntu viewpoint) we need a little more magic before things will work.

As root, create /etc/ld.so.conf.d/libmemcached.conf:

# Manual installation of libmemcached dropped the libs in a non-standard place
/usr/local/lib

Run ldconfig to load the new config and test to ensure your script can now access libmemcached properly through the memcached gem.

Broken memcached gem/libmemchached port on OS X

Update (2009/04/22):
This post was originally only targeted at OS X but I have since noticed Ubuntu also having this same issue. Jump to my additional instructions for Ubuntu.

Recipe for disaster:

1. Use MacPorts for your package management
2. Use the ruby memcached gem
3. Use the libmemcached port as a dependency on the gem above
4. Upgrade your ports

Synopses
I upgraded all my ports on my MBP recently and one of the libs that was upgraded in the process was libmemcached (upgraded from v0.25 to v0.26). Unfortunately this has broken my access to memcached on my system (the script dies when trying to connect).

Not knowing what the problem may be I also upgraded my gem to v0.14 which then could not build. A quick search for the issue found this entry on the memcached discussion forum.

How now brown cow?
Evan Weaver, the maintainer of the memcached project, mentioned that libmemcached v0.26 lacks some critical patch and suggested people may have to wait for v0.27 to rectify things.

For those who cannot wait and have inconsistent, broken systems, Evan goes on to mention that you should currently use the libmemcached lib and memcached gem pair from here.

Here he also provides a convenient compatibility matrix so you can see how to align the different versions of libmemcached and the memcached gem.

Fix
Unfortunately this messes with things because libmemcached now needs to be installed as a source package in a system managed with MacPorts.

Yuck!

Unfortunately the choices are limited as the versions of libmemcached that ships as a port are either v0.26 (the latest, but broken for our purposes) or v0.25 (which aligns with the older v0.13 of the memcached gem).

First off remove any libmemcached ports you may have installed:

$ sudo port uninstall libmemcached

Download the libmemcached source archive, extract, build and install it:

ibm-99tvvxc:msp charl$ cd /var/tmp/
$ wget http://blog.evanweaver.com/files/libmemcached-0.25.14.tar.gz
--2009-02-24 10:50:40--  http://blog.evanweaver.com/files/libmemcached-0.25.14.tar.gz
Resolving blog.evanweaver.com... 208.78.102.192
Connecting to blog.evanweaver.com|208.78.102.192|:80... connected.
HTTP request sent, awaiting response... 200 OK
Length: 428047 (418K) [application/x-gzip]
Saving to: `libmemcached-0.25.14.tar.gz'

100%[==========================================================>] 428,047      135K/s   in 3.1s 

2009-02-24 10:50:49 (135 KB/s) - `libmemcached-0.25.14.tar.gz' saved [428047/428047]

$ tar zxvf libmemcached-0.25.14.tar.gz
libmemcached-0.25.14/
[...]
$ cd libmemcached-0.25.14
$ ./configure
checking build system type... i386-apple-darwin9.6.0
checking host system type... i386-apple-darwin9.6.0
checking target system type... i386-apple-darwin9.6.0
checking for a BSD-compatible install... /usr/bin/install -c
[...]
$ make
Making all in docs
/usr/bin/pod2man -c "libmemcached" -r "" -s 3 libmemcached.pod > libmemcached.3
/usr/bin/pod2man -c "libmemcached" -r "" -s 3 libmemcached_examples.pod > libmemcached_examples.3
[...]
$ sudo make install
Password:
Making install in docs
make[2]: Nothing to be done for `install-exec-am'.
test -z "/usr/local/share/man/man1" || ../config/install-sh -c -d "/usr/local/share/man/man1"
/usr/bin/install -c -m 644 './memcat.1' '/usr/local/share/man/man1/memcat.1'
[...]

You may want to keep the source directory around so that you can easily uninstall (run sudo make uninstall in the source directory) the source installation later when the relevant aligned packaged versions are available for installation.

The final step is to install the gem:

$ sudo env ARCHFLAGS="-arch i386" gem install -r -V memcached --no-rdoc --no-ri
GET 200 OK: http://gems.rubyforge.org/latest_specs.4.8.gz
GET 200 OK: http://gems.github.com/latest_specs.4.8.gz
connection reset after 2 requests, retrying
GET 200 OK: http://gems.rubyforge.org/quick/Marshal.4.8/memcached-0.14.gemspec.rz
Installing gem memcached-0.14
/opt/local/lib/ruby/gems/1.8/gems/memcached-0.14/BENCHMARKS
/opt/local/lib/ruby/gems/1.8/gems/memcached-0.14/CHANGELOG
[...]
Successfully installed memcached-0.14
1 gem installed

Success!

We now have libmemcached v0.25.14 and the memcached v0.14 gem installed and we can go on with our lives.