JRuby on Rails with RVM ( on Fedora )

First install RVM and set it up:

sudo yum install rubygem-rvm
rvm-install

To include RVM into your bash shell, add following line to ~/.bashrc:

[[ -s "$HOME/.rvm/scripts/rvm" ]] && . "$HOME/.rvm/scripts/rvm" # This loads RVM into a shell session.

Now install JRuby and use it as default Ruby interpreter:

source ~/.bashrc
rvm install jruby-1.5.6
rvm use jruby-1.5.6

Setup Rails environment:

gem install rails
gem install bundler

Lets create a new Rails application:

rails new rails-app
cd rails-app
bundle install
rails g scaffold Person name:string
rake db:migrate
rails s

Now go to http://localhost:3000/people/
Yay! Your new Rails application is up and running, powered by JRuby ( setup using RVM ).

Packaging Java JARs as RPM packages

At JUDCon 2012, I had a discussion and argument with Jaikiran Pai and Ravi Maurya on “Packaging Java JARs as RPM packages”.

Well it is not just RPM, it could be any package management system ( eg. dpkg, protage etc. ), doesn’t matter as long as it serves the purpose of:

• automatic dependency resolution
• installation of dependencies
• rollback an install step
• install multiple versions of a (JAR) package
• ensure the packages are authentic ( for security )

How is all this achieved in Java world? Simple. Just package all the required JARs in a single JAR or WAR or an EAR for that matter. And trust the packager for its security. But:

• Is this a good practice?
• What happens when you want to install the same package on multiple systems?
• Can you roll back changes easily on multiple installations?
• How are JARs shared across different kind of projects, some which use the same version of the same JAR package?

Consider the case of Maven, with which you can easily specify the dependencies and Maven does all the work of dependency resolution.
However, it solves the problem only at the developer’s level. JARs are still bundled inside the output application package (JAR / WAR / EAR). And everytime this application is distributed, it will contain “ALL” the dependencies. That is clearly an overhead.

That makes me think, why JARs don’t work like Shared Libraries, which can be shared across applications. Every classloader / JVM instance has its own version of the same JAR file on-disk! Can’t they be shared? That too is clearly an overhead. I am not aware of the reasons why it is so, which I would definitely like to know.

The point is, since a package manager has all the functionality, why not just leverage it? Java community is either unaware of it, or doesn’t care.

No problem. There are efforts already in place.

Anyway, of other things Jaikiran told me that, JBoss AS7 has how better and modular class loading which is based on JBoss Modules. This appears to me similar to how Maven structures it JAR repositories.

Here are some resources:

• What is the proper way to package a single jar file as rpm?
• Packaging Java Apps for Ubuntu @ FOSS.IN
• Java Packaging in Fedora
• Creating an RPM for a Java Application
• Maven plugin to create RPM packages

Thats it folks!

JUDCon 2012 – an experience!

My first Java exclusive conference!

I arrived at the venue right on time. Surprize, there were a lot of people already holding their welcome kits. Conference was already running quite smoothly. Wifi worked most of the time. The sessions, lunch, tea were always on time. Actually, there were people assigned to ring bells, just like in schools, to notify timeout during lunch and tea breaks

The whole confenence was revolving around four major topics: JBoss AS7, Infinispan, Drools and OpenShift.

Opening talk was given by Harish Pillay, Red Hat Global Community Architecture Team.

The Keynote of the day was by Bruno Georges. He mentioned some interesting trends of which are:

• There are about 4 times the processors compared to the number or people in United States alone and growing. Meaning that the processing power is available in abundance.
• There is a significant growth in Ad-Hoc Social Networks for businesses

Now where is the industry heading? In perspective of Java, Java ME/EC is going to be merged with Java SE/EE. SQL based data is moving towards NoSQL which is now leading us to NewSQL. The software stacks are moving towards ubiquitous computing i.e. even smaller handheld devices, are capable of running these software stacks. To sum it up, it was a very factual and a visionary talk by Bruno.

I will defer the other topics till later. Let me talk about the topics I liked the most:

JPA – Painless Persistence

Painless Persistence by Greg Kable was a very engaging talk. It was kind of an OOPS/ORM primer. The concepts were explained in a very general terms like “why/when to use ORM and when not to”, Skinny Objects and their pitfalls. Why to avoid optimization? Such concepts could be applied to any kind of ORM in practice. Even if you didn’t know anything about JPA or Hibernate, you could understand the point Greg was trying to make. And by the way, there are 3 rules of optimization: don’t do it, don’t do it, don’t do it. Before deciding on optimizing part(s) of your application always think of these 3 rules. It is just as a reminder that it might not make sense to put efforts on optimizing your applications first of all.

Future of Seam Framework

Seam 3 is final release of Seam project. That’s it. All future efforts are directed towards Apache DeltaSpike from now on. So, do think twice before using Seam for your new projects. However, Seam being a widely used framework, the bugfixes and mainitainence will still be available. And contributes, checkout the DeltaSpike project now!

Polyglot Revolution

This was an intersting topic. What it means is that many JVM based languages (ceylon, scala, clojure, ruby, python) and their frameworks ( like Ruby On Rails ) will be available integrated with JBoss AS. TorqueBox is a unique example of this. With this effort, you can use all the power and features of Java EE stack right into your other language frameworks. Same argument applies for Immutant.

JBoss Application Server

In the recent years, JBoss has seen a lot of increase in the number of hits on its project sites from the Indian region. Most of these hits are by the users who use JBoss. But what about the size of contributers? What is the plan ahead for community engagement? Well that is the question I asked the panel. JBoss Heroes is project that is supposed to address this issue. Lets see when it is going to actually materialize. JBoss AS7, the latest release, is fully EE6 compliant, is very lightweight and is blazing fast with a boot time less that 3 seconds!

Infinispan

I don’t know why but there were just too many talks on Infinispan alone ( over 6 talks ! ). Inifinispan is a data grid platform based on the design of Amazon Dynamo ( has many implementation including Apache Cassandra ). Infinispan spans a very broad set of usecases. It can act as an in memory cache, persistent cache, distributed cache, replicated cache for high availablity. Also it provides plugins for different backends like Arjuna TM, BDB, JDB, JDBC etc.

Essentially it acts a key/value store. Its real power comes from its flexibility, which allows it to be adapted for many many use cases.

Becoming an Open Source developer

“When in fight, I start to fear that if I get hit by my oppnent, worse is going to happen, and I will lose.” Well, dont! Believe in how good things are going to be if you succeed. Always think that you are going to bring the opponent down no matter what. Never fear of failure. Fear the fear itself, and put your best. That is the mantra of successful Open Source developers, quite well explained by Dmitris in his talk Becoming an Open Source developer. Fight the problem until you solve it. Isn’t that straight forward

Drools – Rules Engine and Complex Event Processing

In the talk “Declarative Rules” by Mark Proctor ( co-creator and founder of Drools ), he explained that Drools is Declarative Rules based engine. He demonstrated a simple game Wumpus World, that can be completely created using only rules. No other code required, which simplifies the whole process. Of the other things in his talk, there qere queries regarding Fuzzy logic support in Drools. Rules reduction or compaction process in Drools. Also rules processing on distributed and unstructured data, which is an inherently hard nut to crack.

In a subsequent talk “Demystifying Complex Event Processing” by Mark Proctor again, he explained how CEP works. How Complex Event Processing (CEP), Event Stream Processing (ESP) and Event Drive Architecture (EDA) are closely related. Drools Fusion is the CEP tool by JBoss. To learn more about Drools / CEP there are many books available: The Power of Events, Event Drive Architecture, Drools JBoss Rules 6.0 Developer’s Guide, JBoss Drools Business Rules

I am happy to have met a lot of people Jaikiran Pai, Sachin Patil, Ray Ploski, Radoslav Husar, Ales Justin, Dmitris Andreadis, and many other whom I only listened to

However, I also missed a few things that would have made this conference even better in my opinion. There could have been an inclusion of activities like Hackathon, Bug Squadding or Workshop ( hands-on sessions ).

It has been a great learning experience. Keep them coming – the events – JUDCon 2012, you rock!

PS: I don’t see any slides nor the videos from JUDCon 2012 on their website yet. When are they coming?

Potential bug in Ruby bindings of Apache Qpid

I wanted to experiment with Ruby bindings of Apache Qpid.

Lets install the the server and client:

$ sudo yum install qpid-cpp-server ruby-qpid qpid-tools
$ service qpidd start

Now check which version of Ruby bindings have we installed here:

$ rpm -qi ruby-qpid
Name        : ruby-qpid
Version     : 0.8
Release     : 2.fc15
Architecture: x86_64
Install Date: Fri 13 Jan 2012 06:21:55 PM IST
Group       : Development/Ruby
Size        : 219877
License     : ASL 2.0
Signature   : RSA/SHA256, Wed 27 Jul 2011 06:41:06 PM IST, Key ID 067f00b6a82ba4b7
Source RPM  : ruby-qpid-0.8-2.fc15.src.rpm
Build Date  : Wed 09 Feb 2011 03:48:16 PM IST
Build Host  : x86-11.phx2.fedoraproject.org
Relocations : (not relocatable)
Packager    : Fedora Project
Vendor      : Fedora Project
URL         : http://qpid.apache.org/
Summary     : Ruby language client for AMQP
Description :
The Apache Qpid project's Ruby language client for AMQP.

Okay, now lets create a simple script and send some messages.

#!/usr/bin/ruby
# file: qpid-test.rb
require 'qpid'

host = "localhost"
port = 5672
address = "amq.topic"
spec_marshal_file = "/usr/lib/ruby/site_ruby/1.8/qpid/spec_cache/amqp.0-10-qpid-errata.rb_marshal"
spec_object = Marshal.load(File.open(spec_marshal_file).read())

socket = TCPSocket.new(host, port)
connection = Qpid::Connection.new(socket, :spec => spec_object)
puts "Connection instance created..."

connection.open

puts "Connection instance opened..."

session    = connection.create_session
puts "Session initiated..."

receiver   = session.create_receiver address
puts "Receiver object created..."

sender     = session.create_sender address
puts "Sender object created..."


sender.send Qpid::Message.new :content => "Hello world!"

message = receiver.fetch Qpid::Messaging::Duration::SECOND
puts "#{message.content}"
session.acknowledge
connection.close

On running the script we should get a “Hello world!” message.

$ ruby qpid-test.rb
Connection instance created...
qpid-test.rb:13qpid-test.rb:13: [BUG] Segmentation fault
ruby 1.8.7 (2011-06-30 patchlevel 352) [x86_64-linux]

Aborted (core dumped)

This is not what we expected. What is wrong here? Perhaps a bug in the Ruby bindings.

Then I went to the source code repository of Ruby bindings on github.com:

https://github.com/apache/qpid/tree/trunk/qpid/cpp/bindings/qpid/ruby

According to the instructions I installed the latest bindings:

$ sudo gem install qpid

And removed the one I installed earlier

$ yum erase ruby-qpid

Now thats funny because the gem I ended up installing using the command is:

http://rubygems.org/gems/qpid

QPID (Queriable Patient Inference Dossier, developed at Massachusetts General Hospital), which is not even related to Apache Qpid.

How am I supposed to installed the latest Ruby bindings for Apache Qpid client?

Symlink loop is still an unsolved problem

How do I dereference a symbolic link?

$ ls -l /usr/bin/java
lrwxrwxrwx. 1 root root 22 Jul 20 2010 /usr/bin/java -> /etc/alternatives/java
$ ls -l /etc/alternatives/java
lrwxrwxrwx. 1 root root 30 Aug 9 2010 /etc/alternatives/java -> /usr/java/jdk1.6.0_20/bin/java
$ ls -l /usr/java/jdk1.6.0_20/bin/java
-rwxr-xr-x. 1 root root 50810 Apr 13 2010 /usr/java/jdk1.6.0_20/bin/java

Lets create a chain of symbolic links:

mkdir /tmp/symlink
cd /tmp/symlink
echo "this is a test file" >  abc.txt
ln -s abc.txt link1
ln -s link1 link2
ln -s link2 link3
ln -s link3 link4

Now lets use the last link that we created:

cat link4
this is a test file

What if we make it a circular link?

rm link1
ln -s link4 link1

Now lets try to use one of the symbolic links:

cat link4
cat: link4: Too many levels of symbolic links

You noticed that? Okay, lets move further:

# we will create a chain of symlinks, which is 50 links deep
mkdir /tmp/links
cd /tmp/links
for i in {0..50}
do
  ln -s l${i} l$(expr $i + 1)
done
echo hello > l0

Lets see what files/links this script has generated:

cd /tmp/links
ls -l

You will notice that all the symlinks upto l40 resolve correctly:

$ cat l40
hello
$ # notice that:
$ # all links from l41 and above will not resolve
$ cat l41 
cat: l41: Too many levels of symbolic links
$ cat l51
cat: l51: Too many levels of symbolic links

Lets print the whole link path now:

# doesn't handle circular symlink loops at the moment
F=l51
while [ -h "$F" ]
do
  G=$(readlink $F)
  echo "$F -> $G"
  F=$G
done
echo "Finally -> $F"

OUTPUT:

l51 -> l50
l50 -> l49
l49 -> l48
...OUTPUT SKIPPED..
l3 -> l2
l2 -> l1
l1 -> l0
Finally -> l0

What’s the bottomline?

• Even if there is no circular symlink loop, the symlink might fail to resolve. This depends on the MAXSYMLINK limit.
• UNIX / Linux doesn’t handle symlink loops at all. The whole assumption is based on a limit. If the symlink chain grows more than a certain limit, we will get a “Too many levels of symbolic links” error, regardless of the symlink loop or not.

References:

http://lwn.net/Articles/91959/

http://www.itworld.com/nls_unix_symboliclinks060622

How to resize a VM and its partitions?

In this post I am going to list the steps to resize a Virtual Machine image created using virt-install

There are three steps:

• Resize the VM image
• Resize the LVM volume inside the VM ( both physical volume and logical volume )
• Resize the filesystem on the LVM volume

First locate where the image for your VM is stored.

# virsh dumpxml vm2 | xpath /domain/devices/disk/source
Found 1 nodes:
-- NODE --
<source file="/export/vmimgs/vm2.img" />

Mine is stored at /export/vmimgs/vm2.img

It is 5GB VM which I want to resize to 10GB. And then, I basically want more space for root partition ( mount point / ).

Step 1: Resize the VM image

Shutdown the VM and take a backup;

# virsh shutdown vm2
# cd /export/vmimgs
# cp vm2.img vm2.img.backup

First lets find the disk devices within our VM:

# virt-filesystems --long -h -a  /export/vmimgs/vm2.img
Name                   Type        VFS   Label  Size
/dev/sda1              filesystem  ext4  -      500M
/dev/VolGroup/lv_root  filesystem  ext4  -      1.5G

# virt-filesystems --long --parts --blkdevs -h -a  /export/vmimgs/vm2.img
Name       Type       MBR  Size  Parent
/dev/sda1  partition  83   500M  /dev/sda
/dev/sda2  partition  8e   1.5G  /dev/sda
/dev/sda   device     -    5G    -

Now we resize our VM image ( +5GB ):

# cd /export/vmimgs
# truncate -r vm2.img vm2.img.new
# truncate -s +5G vm2.img.new 
# virt-resize --expand /dev/sda2 vm2.img vm2.img.new
Examining vm2.img ...
**********

Summary of changes:

/dev/sda1: This partition will be left alone.

/dev/sda2: This partition will be resized from 4.5G to 9.5G.  The LVM 
    PV on /dev/sda2 will be expanded using the 'pvresize' method.

**********
Setting up initial partition table on vm2.img.new ...
Copying /dev/sda1 ...
 100% ⟦▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓⟧ 00:00
Copying /dev/sda2 ...
 100% ⟦▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓⟧ 00:00
 100% ⟦▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓⟧ --:--
Expanding /dev/sda2 using the 'pvresize' method ...

Resize operation completed with no errors.  Before deleting the old 
disk, carefully check that the resized disk boots and works correctly.

Did you notice that /dev/sda2 was resized using ‘pvresize’. Thats because
its a LVM physical volume.

Now we move the new image vm2.img.new to the actuall image, taking backup for safety:

# chmod +x vm2.img.new
# mv vm2.img vm2.img.original
# mv vm2.img.new vm2.img

Step 2: Resize the LVM volume inside the VM – logical volume

Login into guest and resize the logical volume:

[root@localhost ~]# lvresize -L +5G /dev/VolGroup/lv_root

Step 3: Resize the filesystem on the LVM volume.

Resize the ext4 partition:

[root@localhost ~]# resize2fs /dev/VolGroup/lv_root
resize2fs 1.41.12 (17-May-2010)
Filesystem at /dev/VolGroup/lv_root is mounted on /; on-line resizing required
old desc_blocks = 1, new_desc_blocks = 1
Performing an on-line resize of /dev/VolGroup/lv_root to 1708032 (4k) blocks.
The filesystem on /dev/VolGroup/lv_root is now 1708032 blocks long.

[root@localhost ~]# df -h
Filesystem                    Size  Used Avail Use% Mounted on
/dev/mapper/VolGroup-lv_root  6.5G  1.4G  4.8G  23% /
tmpfs                        1004M     0 1004M   0% /dev/shm
/dev/vda1                     485M   31M  429M   7% /boot

Thats it. I hope that helps.

References:

http://libguestfs.org/virt-resize.1.html

http://virt-tools.org/learning/install-with-command-line/

http://docs.redhat.com/docs/en-US/Red_Hat_Enterprise_Linux/6/html/Virtualization/sect-Virtualization-Virtualized_guest_installation_overview-Creating_guests_with_virt_install.html

zif critical errors

Every once in a while I get these errors when installing a package using zif:

Downloading        [==================================================================] (59%)
 packages/glusterfs-server-3.2.4-2.fc16.x86_64.rpm [2.9 MB/s] om 64 to 52 on 0x1fb60b0!
5) zif-main.c:2590 (5/6)
4) zif-main.c:2401 (3/5)
3) zif-transaction.c:3989 (1/3)
2) zif-package-array.c:243 (0/1)
1) zif-package-remote.c:416 (2/3)
0) zif-store-remote.c:505 (1/2)

(zif:18671): Zif-CRITICAL **: percentage should not go down from 60 to 19 on 0x78e6410!
4) zif-main.c:2590 (5/6)
3) zif-main.c:2401 (3/5)
2) zif-transaction.c:3989 (1/3)
1) zif-package-array.c:243 (0/1)
0) zif-package-remote.c:416 (2/3)

(zif:18671): Zif-CRITICAL **: percentage should not go down from 64 to 27 on 0x7f5f4c0!
3) zif-main.c:2590 (5/6)
2) zif-main.c:2401 (3/5)
1) zif-transaction.c:3989 (1/3)
0) zif-package-array.c:243 (0/1)

(zif:18671): Zif-CRITICAL **: percentage should not go down from 64 to 27 on 0x1fb63e0!
2) zif-main.c:2590 (5/6)
1) zif-main.c:2401 (3/5)
0) zif-transaction.c:3989 (1/3)

(zif:18671): Zif-CRITICAL **: percentage should not go down from 61 to 31 on 0x7909ae0!
1) zif-main.c:2590 (5/6)
0) zif-main.c:2401 (3/5)

(zif:18671): Zif-CRITICAL **: percentage should not go down from 49 to 40 on 0x1fb62d0!
0) zif-main.c:2590 (5/6)
Completed          [==================================================================] (100%)
                                                               m 59 to 52 on 0x1fb60b0!
Transaction success!

Although the packages get installed, this is makes me uncomfortable.

And how are these stack-traces generated for C code ?

zif – Package management done faster

On Fedora 16, I used zif. zif is an alternative to yum and its fast!

How to setup zif?
So easy: sudo yum install zif zif-tools

All set!

Search for packages using zif:
sudo zif search mongodb

Do a offline serach ( like cached -C in yum ):
sudo zif search -o mongodb

Install packages using zif:
sudo zif install mongodb-server mysql-server

Update all packages:
sudo zif update

And if you feel like canceling the operation using Ctrl+C, it happens almost instantaneously!

I like it.

How to setup multiple monitors on Xfce desktop?

I just followed the instructionson Xorg RandR 1.2 wiki

http://www.thinkwiki.org/wiki/Xorg_RandR_1.2

Lets find out how many monitors we have connected?

$ xrandr -q
Screen 0: minimum 320 x 200, current 1280 x 800, maximum 8192 x 8192
LVDS1 connected 1280x800+0+0 (normal left inverted right x axis y axis) 261mm x 163mm
   1280x800       60.0*+   50.0  
   1024x768       60.0  
   800x600        60.3     56.2  
   640x480        59.9  
VGA1 connected (normal left inverted right x axis y axis)
   1680x1050      60.0 +
   1280x1024      75.0     60.0  
   1152x864       75.0  
   1024x768       75.1     60.0  
   800x600        75.0     60.3  
   640x480        75.0     60.0  
   720x400        70.1  
HDMI1 disconnected (normal left inverted right x axis y axis)
DP1 disconnected (normal left inverted right x axis y axis)
HDMI2 disconnected (normal left inverted right x axis y axis)
DP2 disconnected (normal left inverted right x axis y axis)
DP3 disconnected (normal left inverted right x axis y axis)

Now ask Xrandr to configure them:

$ xrandr --output LVDS1 --auto --output VGA1 --auto --right-of LVDS1

Now thats simply cool!

Emacs key binding Ctrl+Space not working on Fedora 16 GNOME or XFCE

I installed Fedora 16 x86_64 and then installed Emacs. Then I logged in into (XFCE or GNOME) desktop and started Emacs.
For text selection in Emacs I use Ctrl+Space ( which is the default in Emacs ), and is no longer working. So I cannot select text using Ctr+Space.

However, if I login into virtual terminal ( Ctrl+Alt+F3 ), then login and start Emacs, the key-binding (Ctrl+Space) for text selection works fine. Therefore there is no problem in Emacs itself.

I suspect that the Ctrl+Space key strokes are intercepted by something in between i.e. either GDM or IBus but I am not sure which one. How can I fix it?