Flummoxed by IT

Ruby: Nokogiri installation problem when using Scrubyit

2011-09-06T12:01:00.000-07:00

I have been trying to install Scrubyit and keep getting this error:

$ sudo gem install scrubytERROR: Error installing scrubyt:
webrobots requires nokogiri (>= 1.4.4, runtime)

Logical step is to try to install the webrobots and nokogiri gem.

$ sudo gem install webrobots
$ sudo gem install nokogiri
Building native extensions. This could take a while...
ERROR: Error installing nokogiri:
ERROR: Failed to build gem native extension.
/opt/bitnami/ruby/bin/ruby extconf.rb
checking for libxml/parser.h... yes
checking for libxslt/xslt.h... yes
checking for libexslt/exslt.h... yes
checking for iconv_open() in iconv.h... no
checking for iconv_open() in -liconv... yes
checking for xmlParseDoc() in -lxml2... no
-----
libxml2 is missing. please visit http://nokogiri.org/tutorials/installing_nokogiri.html for help with installing dependencies.

Apparently this is a common enough problem, as this blog post shows. You can install the missing components by running

$ sudo apt-get install libxml-ruby libiconv-ruby

On Mac, you can run

brew install libxml2
brew install libiconv
brew link libiconv

If you are like me and use Bitnami for Ruby development, I suggest using RVM to have better control of the Ruby environment.

Using RVM to control Ruby environment

2011-09-06T11:46:00.000-07:00

I am using a Bitnami Amazon EC2 Rails AMI to do my development work. The paths are rather different

$ gem environment
RubyGems Environment:
- RUBYGEMS VERSION: 1.3.7
- RUBY VERSION: 1.9.2 (2010-12-25 patchlevel 136) [i686-linux]
- INSTALLATION DIRECTORY: /opt/bitnami/ruby/lib/ruby/gems/1.9.1
- RUBY EXECUTABLE: /opt/bitnami/ruby/bin/ruby
- EXECUTABLE DIRECTORY: /opt/bitnami/ruby/bin
- RUBYGEMS PLATFORMS:
- ruby
- x86-linux
- GEM PATHS:
- /opt/bitnami/ruby/lib/ruby/gems/1.9.1
- /home/bitnami/.gem/ruby/1.9.1
- GEM CONFIGURATION:
- :update_sources => true
- :verbose => true
- :benchmark => false
- :backtrace => false
- :bulk_threshold => 1000
- REMOTE SOURCES:
- http://rubygems.org/

Forget about tweaking Ruby paths. Install RVM to help you manage your muliple Ruby environments.

$ bash < <(curl -s https://rvm.beginrescueend.com/install/rvm)

If you install RVM as a regular/single user, run this additional step to ensure RVM is utilized automatically

$ echo '[[ -s "$HOME/.rvm/scripts/rvm" ]] && . "$HOME/.rvm/scripts/rvm" # Load RVM function' >> ~/.bash_profile

Read the RVM installation notes for more details. If you have trouble installing RVM on bitnami AMI, try to use a regular user account, and read this StackOverflow answer.

In fact, I would suggest after installing RVM to avoid using either Bitnami's non-standard Ruby installation, nor use Ubuntu's installation. Instead, install a sandboxed version of Ruby via RVM. Start by listing available Ruby versions

$ rvm list known
# MRI Rubies
[ruby-]1.8.6[-p420]
[ruby-]1.8.6-head
[ruby-]1.8.7[-p352]
[ruby-]1.8.7-head
[ruby-]1.9.1-p378
[ruby-]1.9.1[-p431]
[ruby-]1.9.1-head
[ruby-]1.9.2-p180
[ruby-]1.9.2[-p290]
[ruby-]1.9.2-head
[ruby-]1.9.3[-preview1]
[ruby-]1.9.3-head
ruby-head

The latest stable version at the time of writing is ruby-1.9.2p290. Let's install that.

$ rvm install 1.9.2

Assuming that you have a few versions of Ruby installed, you can ask RVM to list them out.

$ rvm list
rvm rubies
=> ruby-1.9.2-p290 [ i386 ]
ruby-1.8.7-p352 [ i386 ]

To use Ruby 1.9.2

$ rvm use 1.9.2

Using /home/bitnami/.rvm/gems/ruby-1.9.2-p290

$ ruby -v
ruby 1.9.2p290 (2011-07-09 revision 32553) [i686-linux]

Check out this Railcast on RVM

Node.js error ENOTFOUND, Domain name not found

2011-08-06T01:47:00.000-07:00

What does the error ENOTFOUND, Domain name not found means when running node.js?

$ node main.js

events.js:45
throw arguments[1]; // Unhandled 'error' event
^
Error: ENOTFOUND, Domain name not found
at IOWatcher.callback (dns.js:74:15)

A quick search in Google groups shows it is a common issue. It usually means your server cannot bind to the domain name properly. Check your host files. Turned out I got the issue because the developer has modified their host file to listen to "host.local".

var config = {
host : 'local.host',
port : args[2] ||
};
...
...

server.listen(config.port, config.host, function() {
console.log('Express server started on ' + config.host + ':' + config.port);
});

The solution? Just leave out the hostname, and node.js Epress will automatically bind to all ports

server.listen("8080", function() {
console.log('Express server started on ' + config.host + ':' + config.port);
});

Python and Ubuntu

2011-07-30T17:27:00.000-07:00

In my previous post I was complaining that I keep having problems with missing "openid_consumer" Python module. Even rebuilding Django and pip from scratch does not work. So I will try to build Python from scratch now.

Download Python from from either http://www.python.org/download/releases/2.7.1/ or http://www.python.org/getit/releases/3.2/ I am getting the 2.x tarball.

Before you proceed, remove the currently installed python version from Ubuntu.

$ sudo apt-get remove python2.7

Download and extract the tarball.

$ curl -O http://www.python.org/ftp/python/2.7.1/Python-2.7.1.tar.bz2

$ tar xvjf Python-2.7.1.tar.bz2

$ cd Python-2.7.1/

Before you start compiling, make sure you have the GCC compiler installed.

$ sudo apt-get install gcc

Now you are ready to build and install python

$ ./configure && make && sudo make install

The first time I ran it, I got this error:

/usr/bin/python: can't find '__main__' module in '/usr/share/command-not-found'

Turns out that python does not work properly without command-not-found. Go ahead and install it

$ sudo apt-get install command-not-found

The problem with openid_consumer is magically solved, you don't even have to recompile anything.

How to manually install Django and pip

2011-07-30T16:20:00.000-07:00

I have been running the latest version of Ubuntu server, 11.04 Natty to run my Python / Django app. Ran into some problems when I keep getting this error

No module named openid_consumer

Even when I have the Python module django-openid-consumer installed.

I suspect it could be either the Linux environment itself, or some version incompatibilities between Python, Django, and the other modules. I did face some incompatibilities error, especially since my team mate is running Django 1.3 and Ubuntu still ship with 1.2.5

You probably do not want to build Python from scratch, as the default Python version in Ubuntu 11.04 is 2.7.1, the latest in the Python 2.x branch. Unless you really need Python 3.x - not that the latest Django release uses new features only found in Python 3.x

Building Django

Let's start with getting Django's source. Head over to https://www.djangoproject.com/download/, and grab http://www.djangoproject.com/download/1.3/tarball/ . Just in case you are using wget, the file may be downloaded as "index.html"

Resolving media.djangoproject.com... 64.207.133.30
Connecting to media.djangoproject.com|64.207.133.30|:80... connected.
HTTP request sent, awaiting response... 200 OK
Length: 6504003 (6.2M) [application/octet-stream]
Saving to: `index.html'

100%[======================================>] 6,504,003   1.28M/s   in 6.4s

2011-07-30 23:10:52 (998 KB/s) - `index.html' saved [6504003/6504003]

Rename it to django.tar.gz

$ mv index.html django.tar.gz

Extract the contents, then install it - you will need superuser privileges to install to the default locations.

$ tar zxvf django.tar.gz
$ cd Django-1.3/
$ sudo python setup.py install

Building Pip

Pip is a handy utility, similar to Ruby's gems command, that allows you to easily install Python modules with only a single command. Before installing Pip, you need to have virtualenv. Install it this way:

$ curl -O https://raw.github.com/pypa/virtualenv/master/virtualenv.py

$ python virtualenv.py my_new_env

New python executable in potholes/bin/python
Installing setuptools............................done.
Installing pip.....................done.

$ . my_new_env/bin/activate

You can start using pip now:

(my_new_env)$ pip install ...(my_new_env)$ pip install xxx

How to install npm on Ubuntu

2011-07-14T01:07:00.000-07:00

Here is a quick how to for installing npm on Ubuntu, or any variant of UNIX for that matter

$ curl http://npmjs.org/install.sh | sh

It is that simple!

So you want to check out the code from git and compile it yourself? Then follow these steps:

$ git clone https://github.com/isaacs/npm
$ cd npm
$ sudo make install

Now you can install interesting stuff like socket.io

$ npm install socket.io

How to install Node.js on Ubuntu

2011-07-14T00:39:00.000-07:00

Here are short, simple steps for installing node.js and npm on Ubuntu 10.04 There may be slight differences in different version of Ubuntu, but since I am using stock APT command they should work. Check http://help.ubuntu.com for version specific steps.

$ sudo apt-get update
$ sudo apt-get install git-core curl build-essential openssl libssl-dev
$ git clone https://github.com/joyent/node.git

If you get some weird error on SSL certification such as below

error: SSL certificate problem, verify that the CA cert is OK. Details:
error:14090086:SSL routines:SSL3_GET_SERVER_CERTIFICATE:certificate verify failed while accessing https://github.com/joyent/node.git/info/refs

fatal: HTTP request failed

You can learn more about it from http://stackoverflow.com/questions/3777075/https-github-access. I suggest you just bypass it by setting the GIT_SSL_NO_VERIFY variable to true

$ GIT_SSL_NO_VERIFY=true git clone https://github.com/joyent/node.git
$ cd node
$ ./configure
$ make
$ make install

Let's check if node is now installed

$ node -v
v0.5.1-pre

Now let's test it. Create a file called Helloworld.js with the following contents

$ cat helloworld.js
var sys = require('sys')
sys.puts('Hello World')

Then we run it:

$ node helloworld.js
Hello World

That's it!

Set up Vim for Groovy development

2011-07-08T00:16:00.000-07:00

What is the best code editor? I shall leave the question open as everyone has their pet favourite, and usually for good reasons too. It is indisputable that Vim is ubiquitous and lightweight. I frequently only have SSH access to my servers and that is when vim is so much handier compared to other GUI based editors/IDE.

Here's a short guide to set up Vim to edit Groovy:

First, download the latest vim syntax script from http://www.vim.org/scripts/script.php?script_id=945

$ curl http://www.vim.org/scripts/download_script.php?src_id=2926 > ~/.vim/syntax/groovy.vim

Open up /.vim/filetype.vim and add in this line
au BufNewFile,BufRead *.groovy  setf groovy

Open up ~/.vim/scripts.vim and add in these lines to recognize when a groovy file is loaded
if did_filetype()
    finish
  endif
  if getline(1) =~ '^#!.*[/\\]groovy\>'
    setf groovy
  endif

Next, download the groovy plug in that will runs or compiles Groovy scripts, indents code blocks, continues comments on adjacent lines, provides insert-mode abbreviations  from http://www.vim.org/scripts/script.php?script_id=2217

$ curl http://www.vim.org/scripts/download_script.php?src_id=8600 > ~/.vim/ftplugin/groovy.vim

That's it! Now when you open up a groovy file in Vim, the syntax will be highlighted and you get auto-indents.

Synchronizing PandaForm with Google Spreadsheets

2011-07-03T09:35:00.000-07:00

So now you have the survey results back from your participants. Congratulations! That's a big first step to understand what the customers really want. Now PandaForm is useful to create forms and surveys, but it is no place for doing in depth analysis.

To do that, you can use the free and very powerful Google Spreadsheets. First, open up the survey form, click on Export and select Google Spreadsheet

Next, select the option Export all and then proceed to Log in to your Google Account

Once you have keyed in your username (the same as your Gmail username if you have one) and password, you will be asked to grant the permissions to allow PandaForm to export to your Google Docs account. Click on Grant access to continue.

Once you have granted access, click on Create a new spreadsheet. You can actually choose to export to an existing spreadsheet as well. Do not worry about the Sheet name, as you can choose it later, or export it to a brand new sheet.

Type in the new spreadsheet name. Use a meaningful name that can be easily understood, because are you create more and more documents in your Google Docs account, trying to figure out the contents of "Survey" or "Survey 5" will be a pain.

Next, type in a new sheet name. I strongly suggest you use a meaningful sheet name such as "Survey" or "Results" instead of the default "Sheet 1".

Select the option to automatically sync the data from Pandaform to Google Spreadsheet if you expect further updates.

After you have clicked Export, you will see a message informing you the export is happening in the background. You can continue performing other tasks in PandaForm

When you now log in to your Google Docs account, you will be able to see the survey results available in a Google Spreadsheet.

Formspring is asking for way too much information!

2011-06-27T13:41:00.000-07:00

Formspring is a promising new startup where you can create profile pages and allow anonymous users to ask questions. This simple idea has generated 25 million users in the span of 18 months.

What I do not like is the fact that it is asking for way too much information. Take for example for sign up, you can use your Facebook account. However, it is still asking for your email as well as date of birth as mandatory information, without explaining what it will be doing with.

Visibli causing duplicate Tweets

2011-06-27T12:57:00.000-07:00

Visible is a pretty neat tool to track engagement of your brand, company or even yourself in both Facebook and Twitter

Unfortunately after I signed up, it seems to automatically double-post my Tweets

Anyone knows of a fix?

Generating total number of requirements for each Requisite Pro projects

2011-06-09T04:56:00.000-07:00

Have you ever wanted to have a summary of all Requisite Pro projects and their total number of requirements?

Run this SQL Server Query / View to generate the output:

USE [RequisitePro]
GO
/****** Object:  View [ReqPro].[All_Requirements_Extract]    Script Date: 06/09/2011 13:03:33 ******/
SET ANSI_NULLS ON
GO
SET QUOTED_IDENTIFIER ON
GO
ALTER view [RePro].[All_Requirements_Extract] as
select p.name ProjectName, t.name ReqTypeName, count(*) Total
from ReqPro.rqprojects p , ReqPro.rqprojectrequirementtypes a, ReqPro.rqrequirementtypes t, ReqPro.rqrequirements r
where p.id = a.projectid
and t.id = a.requirementtypeid
and r.requirementtypeid = a.requirementtypeid
group by p.name, t.name

You should see 3 columns: Requisite Pro project name, the requirement type and the count

Workaround for Bitnami's custom PATH

2011-06-05T13:25:00.000-07:00

Have you ever face error with path, especially when compiling an application from source in a Bitnami EC2 instance?

For example, I get this error while trying to compile the s3fs program:

configure: error: Package requirements (fuse >= 2.8.4 libcurl >= 7.0 libxml-2.0 >= 2.6 libcrypto >= 0.9) were not met:

No package 'fuse' found
No package 'libcurl' found
No package 'libxml-2.0' found
No package 'libcrypto' found

Consider adjusting the PKG_CONFIG_PATH environment variable if you
installed software in a non-standard prefix.

Alternatively, you may set the environment variables DEPS_CFLAGS
and DEPS_LIBS to avoid the need to call pkg-config.
See the pkg-config man page for more details.

A check in $PATH shows that a lot of /opt/bitnami/... directories is added to the front of PATH

# echo $PATH

/opt/bitnami/memcached/bin:/opt/bitnami/perl/bin:/opt/bitnami/git/bin:/opt/bitnami/nginx/sbin:/opt/bitnami/sphinx/bin:/opt/bitnami/sqlite/bin:/opt/bitnami/php/bin:/opt/bitnami/mysql/bin:/opt/bitnami/apache2/bin:/opt/bitnami/subversion/bin:/opt/bitnami/ruby/bin:/opt/bitnami/common/bin:/opt/bitnami/memcached/bin:/opt/bitnami/perl/bin:/opt/bitnami/git/bin:/opt/bitnami/nginx/sbin:/opt/bitnami/sphinx/bin:/opt/bitnami/sqlite/bin:/opt/bitnami/php/bin:/opt/bitnami/mysql/bin:/opt/bitnami/apache2/bin:/opt/bitnami/subversion/bin:/opt/bitnami/ruby/bin:/opt/bitnami/common/bin:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games

Not smart, if you want to use stock Debian packages.

Solution? Re-export the PATH to contain the default paths first:

# export PATH=/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:$PATH

By the way, if its s3fs that you need to compile, you need to the development libraries, including libcurl:

# apt-get install libcurl4-openssl-dev

Adventures installing google-api-client on Bitnami EC2 instance

2011-06-05T10:16:00.000-07:00

I launched an AWS EC2 AMI ami-d09be582 Bitnami Rails stack to try out Google's Ruby API.

~$ sudo gem install google-api-client

Building native extensions. This could take a while...

ERROR: Error installing google-api-client:

ERROR: Failed to build gem native extension.

/opt/bitnami/ruby/bin/ruby extconf.rb

checking for ruby/re.h... *** extconf.rb failed ***

Could not create Makefile due to some reason, probably lack of

necessary libraries and/or headers. Check the mkmf.log file for more

details. You may need configuration options.

Provided configuration options:

--with-opt-dir

--without-opt-dir

--with-opt-include

--without-opt-include=${opt-dir}/include

--with-opt-lib

--without-opt-lib=${opt-dir}/lib

--with-make-prog

--without-make-prog

--srcdir=.

--curdir

--ruby=/opt/bitnami/ruby/bin/ruby

/opt/bitnami/ruby/lib/ruby/1.9.1/mkmf.rb:368:in `try_do': The complier failed to generate an executable file. (RuntimeError)
You have to install development tools first.
from /opt/bitnami/ruby/lib/ruby/1.9.1/mkmf.rb:452:in `try_cpp'
from /opt/bitnami/ruby/lib/ruby/1.9.1/mkmf.rb:834:in `block in have_header'
from /opt/bitnami/ruby/lib/ruby/1.9.1/mkmf.rb:693:in `block in checking_for'
from /opt/bitnami/ruby/lib/ruby/1.9.1/mkmf.rb:280:in `block (2 levels) in postpone'
from /opt/bitnami/ruby/lib/ruby/1.9.1/mkmf.rb:254:in `open'
from /opt/bitnami/ruby/lib/ruby/1.9.1/mkmf.rb:280:in `block in postpone'
from /opt/bitnami/ruby/lib/ruby/1.9.1/mkmf.rb:254:in `open'
from /opt/bitnami/ruby/lib/ruby/1.9.1/mkmf.rb:276:in `postpone'
from /opt/bitnami/ruby/lib/ruby/1.9.1/mkmf.rb:692:in `checking_for'
from /opt/bitnami/ruby/lib/ruby/1.9.1/mkmf.rb:833:in `have_header'
from extconf.rb:17:in `

Gem files will remain installed in /opt/bitnami/ruby/lib/ruby/gems/1.9.1/gems/json-1.5.1 for inspection.
Results logged to /opt/bitnami/ruby/lib/ruby/gems/1.9.1/gems/json-1.5.1/ext/json/ext/generator/gem_make.out

Just install ruby-dev package, simple enough to resolve right? Wrong

bitnami@ip-10-136-78-35:~$ sudo apt-get update && apt-get install ruby1.9-dev

The following extra packages will be installed:

binutils gcc gcc-4.4 libc-bin libc-dev-bin libc6 libc6-dev libc6-i686

libc6-xen libgomp1 libruby1.9 linux-libc-dev manpages-dev ruby1.9

Suggested packages:

binutils-doc gcc-multilib autoconf automake1.9 libtool flex bison gdb

gcc-doc gcc-4.4-multilib libmudflap0-4.4-dev gcc-4.4-doc gcc-4.4-locales

libgcc1-dbg libgomp1-dbg libmudflap0-dbg libcloog-ppl0 libppl-c2 libppl7

glibc-doc ruby1.9-examples rdoc1.9 ri1.9 rubygems1.9

The following NEW packages will be installed:

binutils gcc gcc-4.4 libc-dev-bin libc6-dev libgomp1 libruby1.9

linux-libc-dev manpages-dev ruby1.9 ruby1.9-dev

The following packages will be upgraded:

libc-bin libc6 libc6-i686 libc6-xen

4 upgraded, 11 newly installed, 0 to remove and 117 not upgraded.

Need to get 25.7MB of archives.

After this operation, 62.8MB of additional disk space will be used.

Do you want to continue [Y/n]?

Installing ruby-dev will update libc6, certainly something that you don't want to do, because you risk further binary compatibility problems.

The weird thing is if you attempt to install ruby-dev package again, apt will attempt to install ruby1.8-dev!

bitnami@ip-10-136-78-35:~$ sudo apt-get install ruby-dev

Reading package lists... Done

Building dependency tree

Reading state information... Done

The following extra packages will be installed:

libruby1.8 ruby1.8 ruby1.8-dev

Suggested packages:

ruby1.8-examples rdoc1.8 ri1.8

The following NEW packages will be installed:

libruby1.8 ruby-dev ruby1.8 ruby1.8-dev

0 upgraded, 4 newly installed, 0 to remove and 117 not upgraded.

Need to get 2,058kB of archives.

After this operation, 7,885kB of additional disk space will be used.

Do you want to continue [Y/n]?

Anyway, you should be able to install google-api-client now

bitnami@ip-10-136-78-35:~$ sudo gem install google-api-client

Building native extensions. This could take a while...

Successfully installed json-1.5.1

Successfully installed signet-0.2.2

Successfully installed extlib-0.9.15

Successfully installed configuration-1.2.0

Successfully installed launchy-0.4.0

Successfully installed liquid-2.2.2

Successfully installed google-api-client-0.2.0

7 gems installed

Installing ri documentation for json-1.5.1...

Installing ri documentation for signet-0.2.2...

Installing ri documentation for extlib-0.9.15...

Installing ri documentation for configuration-1.2.0...

Installing ri documentation for launchy-0.4.0...

Installing ri documentation for liquid-2.2.2...

Installing ri documentation for google-api-client-0.2.0...

Installing RDoc documentation for json-1.5.1...

Installing RDoc documentation for signet-0.2.2...

Installing RDoc documentation for extlib-0.9.15...

Installing RDoc documentation for configuration-1.2.0...

Installing RDoc documentation for launchy-0.4.0...

Installing RDoc documentation for liquid-2.2.2...

Installing RDoc documentation for google-api-client-0.2.0...

Another problem that I encournter is that apt refuses to install ruby-dev, complaining about paths to /opt/bitnami where Rails lives, instead of using the standard path in Debian/Ubuntu. Since I already have the debian package locally, I wanted to extract it using ar. ar is part of binutils, I installed binutils which result in libc6 updated. Crazy! Thankfully its a throwaway EC2 Ubuntu instance.

I can see why people who use Heroku don't turn back.

How to sync Phone and SIM contacts with Google on Android phone

2011-05-31T03:42:00.000-07:00

I just realized that Phone contacts and Google contacts are 2 totally different groups, and you cannot easily sync in Contacts app. I am using Samsung Galaxy S, but there should be minimal differences between Android phones. Follow these steps:

1. Go to Contacts -> Menu -> Display Options.

2. Select only Phone and make sure you deselect the other lists such as SIM card, Google, Facebook, Twitter, etc

3. Click on Menu -> More -> Import / Export and select Export to SD Card You should see a 0xxxxx.vcf file being created in /sdcard/

4. You should be back to the Contacts main screen. Click on Menu -> Import/Export and this time select Import from SD Card. You should select Google, for example first.last@google.com

5. Once the import process is completed, go to Menu -> Display Options again. This time, de-select Phone and select your Google accounts, as well as any other online accounts.

I make it a point to only store contacts to Google so that I can easily manage it in one place, and make a backup of my complete contact list from time to time. This way, even if I lose my phone, use another SIM card or go to another country, I will have access to my contact list as long as I am using an Android phone.

Cannot hibernate in Windows Vista

2011-04-25T04:51:00.000-07:00

Are you trying to hibernate your Windows Vista machine, and the "Hibernate" option is missing?

Run the command "powercfg /hibernate on" and the Hibernate option will be available again

Eclipse Android error - An SDK Target must be specified

2011-04-19T11:54:00.000-07:00

When you create your first Android project in Eclipse
it is common to see this error message: "An SDK target must be specified"

The Target Name option is disabled.

The error really means that Eclipse cannot find the installation directory of Android SDK. Go to Window -> Preferences. Click on Android on the left navigation pane, then browse for the installation directory of Android SDK.

Click on Apply, and you will see the list of available SDK target versions. Select an appropriate version, then click on OK

Now when you create a new Android project, you will be able to select an appropriate SDK version and proceed with project creation.

Vista poor performance after installing patches

2011-04-19T03:16:00.000-07:00

I always experience Vista performing very slowly, or not booting up properly after installing patches. Anyone else have the same issue?

tasksel error

2011-03-27T15:41:00.000-07:00

I just spin up an EC2 Ubuntu server image and trying to install the LAMP stack, when I got this error:

$ sudo tasksel install lamp-server

tasksel: aptitude failed (100)
Googled around and did not immediately found a solution. There was a suggestion to use:

$ sudo dpkg -configure -a

which did not work for me.

Tried this and worked:

$ sudo apt-get update

IIS .NET compatibility error: Child nodes not allowed

2010-12-06T01:31:00.000-08:00

I got this weird error message from IIS 6.0:

Parser Error Message: Child nodes not allowed.

The web.config line causing this error is:

The error message does not exactly tell you about incompatible .NET version; in my case I will need to have .NET 3.5 installed and IIS configured. Download .NET Framework 3.5 here and install it, IIS should automatically be able to refer to .NET 3.5 libraries.

Note that if you right-click on a Web Site in IIS Manager, select Properties and go to the ASP.NET tab, you may be surprised to still see reference to .NET 2.0

For more information of the differences between .NET Framework 2.0, 3.0 and 3.5, refer to this article http://www.hanselman.com/blog/HowToSetAnIISApplicationOrAppPoolToUseASPNET35RatherThan20.aspx

This diagram sums it up nicely:

yforth - is this a joke?

2010-10-15T22:03:00.000-07:00

Guess what I get when I search for "joke" using apt-cache?

hanxue@linuxmint ~ $ apt-cache search joke
yforth - A small freeware Forth environment in ANSI C

bash script to query AWS EC2 meta-data

2010-10-14T09:54:00.000-07:00

Do you know that you can actually query EC2 meta-data by making http request to http://169.254.169.254/latest/user-data . For more information, please see http://developer.amazonwebservices.com/connect/entry.jspa?externalID=1825

Get the script from:

http://s3.amazonaws.com/ec2metadata/ec2-metadata

Add it to your PATH and make it executable. Then you can run it:

$ ec2-metadata
ami-id: ami-381e606a
ami-launch-index: 0
ami-manifest-path: (unknown)
ancestor-ami-ids: not available
block-device-mapping:
ami: /dev/sda1
ephemeral0: /dev/sda2
ephemeral1: /dev/sda3
root: /dev/sda1
instance-id: i-28fcf97a
instance-type: t1.micro
local-hostname: ip-x-x-x-x.ap-southeast-1.compute.internal
local-ipv4: x.x.x.x
kernel-id: aki-bdf38cef
placement: ap-southeast-1a
product-codes: not available

Ignore GPG Key authentication when installing from Debian repository

2010-10-14T09:22:00.000-07:00

I am trying to install x2go on an Amazon AWS EC2 instance, but for some reason I fail to import the gpg keys using the command:

gpg --keyserver wwwkeys.eu.pgp.net --recv-keys C509840B96F89133

Since this is a once-off usage of the Debian repository, and there is no valuable data being risked, I opted to ignore GPG key authentication. To tell APT to ignore GPG, edit the file /etc/apt/apt.conf.d/01ubuntu to this:

APT
{Ignore {"gpg-pubkey"; }
};

You will be warned that some of the Debian packages cannot be authenticated. Answer y to proceed:

WARNING: The following packages cannot be authenticated!
  nxproxy pinentry-x2go x2goclient
Install these packages without verification [y/N]? y

Apache and Tomcat

2010-08-26T13:56:00.000-07:00

I believe it is quite common to configure Tomcat and Apache to run on the same server. There are various reasons to do so, mainly for performance reasons, because Apache is much more efficient at serving static content, compared to Tomcat.

My reason is because I have a PHP application running on Apache, and I wanted to run another J2EE application. For some reason, I fail to open up the 8080 port to the public - its an Amazon AWS EC2 instance and I don'd intend to write more on AWS Security Groups at this point. To cut to the chase, I want to redirect request using a specific path, such as http://domain.com/j2ee to localhost:8080 where Tomcat is running.

There are 2 methods, the simpler being using mod_proxy in Apache.

mod_proxy

1. Enable mod_proxy. You don't have to install anything as mod_proxy is installed by default in Ubuntu/Debian
# a2enmod proxy

2. Edit Apache configuration at /etc/apache2/apache2.conf

ProxyPass /alfresco http://localhost:8080/
ProxyPassReverse /alfresco http://localhost:8080/

3. Restart Apache

# service apache2 restart

AJP and mod_jk

1. Install mod_jk for Apache

# apt-get install libapache2-mod-jk

This package contains a sample Apache2 configuration located at /usr/share/doc/libapache2-mod-jk/httpd_example_apache2.conf

There is already a sample worker.properties file at /etc/libapache2-mod-jk/workers.properties

2. Assuming you want to accept the default option, you can just append the contents of httpd_example_apache2.conf into the existing apache2.conf

# cat /usr/share/doc/libapache2-mod-jk/httpd_example_apache2.conf >> /etc/apache2/apache2.conf

3. You need to reload the new Apache2 configuration

# service apache2 reload

4. List all the modules loaded by Apache, and make sure mod_jk is loaded too

# apache2 -t -D DUMP_MODULES

You should see this line

jk_module (shared)

In case you get this error:

apache2: bad user name ${APACHE_RUN_USER}

you may need to edit /etc/apache2/apache2.conf and change them from

User ${APACHE_RUN_USER}
Group ${APACHE_RUN_GROUP}

to the static values:

User www-data
Group www-data

5. Let's take a look at the worker.properties file. You need to ensure the workers.tomcat_home variable is set correctly. By default it is workers.tomcat_home=/usr/share/tomcat5

Find out which process is using the most CPU

2010-08-26T12:21:00.000-07:00

I had Linux running on a small EC2 Amazon AWS instance, and out of nowhere the system load is averaging at 1.0

Surprisingly top only shows system processes, nothing that indicate the load is out of normal

PID USER PR NI VIRT RES SHR S %CPU %MEM TIME+ COMMAND
1 root 20 0 2808 1536 1204 S 0.0 0.1 0:00.13 init
2 root 20 0 0 0 0 S 0.0 0.0 0:00.00 kthreadd
3 root RT 0 0 0 0 S 0.0 0.0 0:00.00 migration/0
4 root 20 0 0 0 0 S 0.0 0.0 0:00.00 ksoftirqd/0
5 root RT 0 0 0 0 S 0.0 0.0 0:00.00 watchdog/0
6 root 20 0 0 0 0 S 0.0 0.0 0:00.01 events/0
7 root 20 0 0 0 0 S 0.0 0.0 0:00.00 cpuset

I guess I was used to the older version of top where the 'C' keystroke will automatically sort by highest CPU usage.

Turns out the default sorts by %CPU usage, which may not accurately show you the process that sucks up most of the current CPU time. Press 'K' to select column to sort, then 'L' to sort by CPU time

Current Sort Field: L for window 1:Def
Select sort field via field letter, type any other key to return

a: PID = Process Id v: nDRT = Dirty Pages count
b: PPID = Parent Process Pid w: S = Process Status
c: RUSER = Real user name x: COMMAND = Command name/line
d: UID = User Id y: WCHAN = Sleeping in Function
e: USER = User Name z: Flags = Task Flags
f: GROUP = Group Name
g: TTY = Controlling Tty Note1:
h: PR = Priority If a selected sort field can't be
i: NI = Nice value shown due to screen width or your
j: P = Last used cpu (SMP) field order, the '<' and '>' keys
k: %CPU = CPU usage will be unavailable until a field
* L: TIME = CPU Time within viewable range is chosen.

Ahh... now it shows the culprit is a Java application:

PID USER PR NI VIRT RES SHR S %CPU %MEM TIME+ TIME COMMAND
16882 root 20 0 1419m 344m 9740 S 0.0 20.3 0:31.90 0:31 /usr/bin/jav
8646 www-data 20 0 52800 25m 4184 S 0.0 1.5 0:28.28 0:28 /usr/sbin/ap
12433 www-data 20 0 55880 28m 4324 S 0.0 1.7 0:22.95 0:22 /usr/sbin/ap
6455 mysql 20 0 149m 36m 6796 S 0.0 2.1 0:12.75 0:12 /usr/sbin/my
13535 www-data 20 0 48816 20m 3840 S 0.0 1.2 0:03.22 0:03 /usr/sbin/ap

Anyway, a bit of googling shows that this command will display the 10 highest CPU grossing processes:

# ps -eo pcpu,pid,user,args | sort -k 1 -r | head -10
%CPU PID USER COMMAND
0.2 16882 root /usr/bin/java -Djava.util.logging.config.file=/opt/alfresco/tomcat/conf/logging.properties -Xms512m -Xmx1024m -Xss1024k -XX:MaxPermSize=256m -XX:NewSize=256m -server -Dalfresco.home=. -Dcom.sun.management.jmxremote -Djava.util.logging.manager=org.apache.juli.ClassLoaderLogManager -Djava.endorsed.dirs=/opt/alfresco/tomcat/endorsed -classpath /opt/alfresco/tomcat/bin/bootstrap.jar -Dcatalina.base=/opt/alfresco/tomcat -Dcatalina.home=/opt/alfresco/tomcat -Djava.io.tmpdir=/opt/alfresco/tomcat/temp org.apache.catalina.startup.Bootstrap start
0.0 9 root [netns]
0.0 8 root [khelper]
0.0 8646 www-data /usr/sbin/apache2 -k start
0.0 7 root [cpuset]
0.0 7904 root /bin/bash
0.0 758 root /bin/bash
0.0 757 root SCREEN
0.0 7375 root /usr/sbin/apache2 -k start

Further explanation of the ps flags:

Flag

Flag	Purpose
e	All processes in the system
o	formatting - show the following fields : pcpu,pid,user,args

The sort command will move processes with highest CPU time (pcpu, column 1) to the top. Head obviously shows the top 10 results.

Further information on troubleshooting performance:

http://www.cyberciti.biz/tips/how-do-i-find-out-linux-cpu-utilization.html

http://www.cyberciti.biz/faq/linux-performance-tools-to-troubleshoot-problem/

How to terminate processes from command line in Windows 2008

2010-08-18T07:49:00.000-07:00

I am connected to a Citrix session, and wanted to terminate a process that I have started up. This means that:

I do not have administrator rights to the server
I do not even have full desktop access, only via a Citrix-streamed application

So the usual method of launching the Task Manager, then use End Process, does not work.

In Windows 2008, there are a number of useful commands, including 2 that can list out all processes and to terminate them.

To list out all running tasks:

C:\> tasklist

Image Name                     PID Session Name        Session#    Mem Usage
========================= ======== ================ =========== ============
System Idle Process              0                            0         28 K
System                           4                            0        244 K
smss.exe                       548                            0        516 K
csrss.exe                      596                            0      4,824 K
winlogon.exe                   624                            0     24,932 K
services.exe                   672                            0     10,600 K
lsass.exe                      684                            0     35,812 K
svchost.exe                    888                            0      4,964 K
svchost.exe                    972                            0      6,368 K
svchost.exe                   1044                            0      6,432 K
svchost.exe                   1096                            0      8,304 K

To terminate them, use taskkill . There are a few ways to do it, run taskkill /? for the available options. In UNIX/Linux, you need to know the process ID (PID) to terminate the process. Using taskkill, you can even specify the process/image name:

c:\>taskkill /IM taskmgr.exe
SUCCESS: Sent termination signal to the process "taskmgr.exe" with PID 19524.
SUCCESS: Sent termination signal to the process "taskmgr.exe" with PID 20452.

Of course, using PID is possible:

C:\>taskkill /PID 19254

To force terminate a process, use the /F flag. I.e.

C>\>taskkill /F /PID 19254

Manually start Language Bar in Windows Safe Mode; Putty behind firewall

2010-08-17T04:58:00.000-07:00

I have been facing frequent crashes in Windows Vista in the past 2 weeks, so decided to start up Windows in safe mode and to do some maintenance - the usual file deletions, uninstall of unneeded applications, disk defragmentation, compression of Outlook PST.

File deletions - Windows Disk Cleanup is sufficient for most purposes. Chances are most of the disk space is taken up by files that you have downloaded/created. For manual clean up:

delete any unsued files
empty Recycle Bin
Remove older System Restore versions - %systemroot%\system32\rstrui.exe
Delete files from the Windows temporary folder in %windir%\temp , as well as your own %temp% folder. Usually it is located in C:\Users\\AppData\Local\Temp

Uninstall applications - run "appwiz.cpl" and uninstall anything that you no longer require

Disk Defragmentation - You know you need to run this frequently on any Windows systems. The basic disk defragmenter that ships with Windows does its basic function well. In fact, it is scheduled to run in the background, without any GUI.

A better option is Auslogic's Disk Defragmenter. It runs a few times faster than Windows disk defragmenter, and seems to be doing its job more effectively too. Remember to close any application and terminate unneeded services to have maximum effect. Alternatively, run Disk Defragmenter in Safe Mode.

Compress Outlook Data Files - The main outlook email "folder" is stored C:\Users\\AppData\Local\Microsoft\Outlook\outlook.ost . If you create additional data files, then you know the location of the files with .pst extension.

Ideally these files should be less than 2 GB in size. As emails are moved/deleted, the deleted items still take up actual storage space. By compacting, the unneeded space is recovered. The detailed steps for compacting Outlook data files is in this Microsoft Knowledge base article -http://support.microsoft.com/kb/289987

Summary of how I do it on Microsoft Outlook 2007 in Windows Vista SP1:

Compacting Inbox / outlook.ost

Delete any items that you do not want to keep, and then empty the Deleted Items folder.
On the Tools menu, click Account Settings.
In the list, select Microsoft Exchange Server, and then click Change.
Click More Settings.
On the Advanced tab, click Offline Folder File Settings.
Click Compact Now.

Compacting PST

On the File menu, click Data File Management.
Click to select your Personal Folder, and then click Settings.
On the General tab, click Compact Now.
Click OK, and then click Close.

Back to the problem I am facing. In Windows Vista's safe mode, I cannot get these to be working:

Office Communicator 2007 cannot run because it fails to subscribe to Windows Event Notification Service
Firewall agent not working - therefore not possible to ssh to external machines using putty.exe
Accidentally activated Asian Language. In Control Panel, a lot of options including Regional Option is not available

So what are the solutions?

Manuallly starting Language Bar - Just go to Start -> Run and type in "cftmon"

SSH to external host -set up Putty to use my company's proxy server:

SQL Server Myths

2010-04-11T21:16:00.000-07:00

See them all at http://www.sqlskills.com/BLOGS/PAUL/

Data file shrink will decrease performance by causing index fragmentation
Enabling AWE in 64-bit does nothing
You can only have a single mirror, but multiple log shipping is possible. Info on threads for mirroring. Microsoft KB - things to consider when setting up mirroring.
DLL triggers are NOT implemented as INSTEAD OF triggers
Database mirroring does not detect failure immediate

Principal SQL Server instance dies - 1 second
OS of principal SQL server dies - 10 seconds
Log drive unavailable - 20 seconds

Online index operations does acquire lock
The null bitmap is always present, contains bits for all columns, adding a column does not result in immediate size-of operation
You cannot enable instant file initialization within SQL Server, but can temporarily disable it by turning on trace flag 1806. More details here.
DBCC CheckDB is lock-free
After failover, any in-flight transaction will NOT be recovered

Querying SQL Server backup results

2010-04-11T20:13:00.000-07:00

Here is a simple query to list out the physical device name of your database backup

  SELECT          physical_device_name,
                          backup_start_date,
                          backup_finish_date,
                          backup_size/1024.0 AS BackupSizeKB
   FROM msdb.dbo.backupset b
   JOIN msdb.dbo.backupmediafamily m ON b.media_set_id = m.media_set_id
   WHERE database_name = 'CF_WIP'
   ORDER BY backup_finish_date DESC

You can also list out the most recent database backup, sorted by backup completion date/time

WITH LastBackupTaken AS (
SELECT database_name,
backup_finish_date,
RowNumber = ROW_NUMBER() OVER (PARTITION BY database_name
ORDER BY backup_finish_date DESC)
FROM msdb.dbo.backupset

)

SELECT database_name,backup_finish_date
FROM LastBackupTaken
WHERE RowNumber = 1

Further explanation at http://blogs.lessthandot.com/index.php/DataMgmt/DBAdmin/MSSQLServerAdmin/getting-the-physical-device-name-and-bac

Reset Terminal Service Client license

2010-03-22T17:19:00.000-07:00

Have you ever had trouble connecting to a remote server via Microsoft Terminal Services Client? Where only a specific server refuse to connect you, but other servers work fine?

It may be due to expiry of client-side terminal services license.

To reset it:

1. Open up the Registry Editor (Start -> Run -> "regedit.exe")
2. Navigate to HKEY_LOCAL_MACHINE\SOFTWARE\MICROSOFT\MSLicensing\HardwareID . Backup the key in case anything goes wrong
3. Delete the Binary value ClientHWID
4. Start up Terminal Services client again. Your license should be refreshed from the server

How to open .jar files as an archive folder?

2010-02-25T00:07:00.000-08:00

Since Windows XP and Windows Server 2003, it is possible to open .zip file as a folder, without relying external applications such as Winzip. However, Java-related archives such as .jar, .ear and .war does not open by default.

Simple trick to get Windows to recognize it as a compressed folder:

assoc .jar=CompressedFolder

If that somehow does not work, it is worth trying a more esoteric solution:

1. Open Registry Editor and export the HKEY_CLASSES_ROOT\.zip key
2. Edit the .reg file and change all references of .zip to .jar
3. Import it to HKEY_CLASSES_ROOT\.jar

Good luck!

Email Stubbing

2010-02-04T11:35:00.000-08:00

What it is: moving attachment or whole message body from mail store to an archive
Advantages:

Reduce mailbox size
Ability to store much more emails in user's mailbox
Message retention policy

Disadvantages:

Search problems (slow, incomplete)
Performance

Bottom line: If you are using Microsoft Exchange / Outlook, don't enable email stubbing

Further information:
http://www.ferris.com/?p=320762
http://technet.microsoft.com/en-us/library/cc671168%28EXCHG.80%29.aspx

Firefox and Windows NTLM/Kerberos authentication

2010-02-04T08:47:00.000-08:00

My company is a primarily Microsoft/Windows-based shop, though I used to do UNIX/Linux support. There are plenty of reasons I prefer using Firefox over Internet Explorer, but unfortunately I still use IE when it comes to intranet websites due to Single Sign On (SSO) requirements.

Finally figure out how to make Firefox work with intranet website using NTLM/Kerberos authentication. Just need to turn these options on in Firefox:

network. ntlm. send-lm-response = true
network. negotiate-auth. using-native-gsslib = true
network. negotiate-auth. trusted-uris = https://, http://
network. negotiate-auth. delegation-uris = https://, http://

Oh, did I mention that you need to go to the Address Bar in Firefox and type in "about:config" first?

Internet Explorer Proxy Settings - Part 2

2009-11-11T11:58:00.001-08:00

I have recently faced problems accessing an intranet website when connected through VPN. The website at http://citrix.company.com will attempt to perform Kerberos authentication at the fowarded URL http://citrix.company.com/XenApp/auth/login.aspx

On further investigation, the URL works perfectly fine when accessed on a machine directly in the company LAN. These are the steps you can take to resolve the problem

Internet Explorer Proxy Settings
Ensure the correct proxy settings is set up. If you are connected through ISA server with PAC, then use "Automatically detect settings". Otherwise fill in the correct proxy values for the VPN connection

Internet Explorer Proxy Settings

If the proxy settings still does not help, it may be due to Trust Settings in Internet Explorer.

1. Go to Tools -> Internet Options -> Security
2. Click on "Local Intranet" and disable protected mode
3. Click on Sites -> Advanced then add in a wildcard entry for the site to override or pass through proxy server

Registry Settings

If you cannot edit the security settings, you can try tweaking the Registry settings.

1. Start up Registry Editor and go to HKEY_LOCAL_MACHINE\Software\Policies\Microsoft\Windows\CurrentVersion\Internet Settings

2. Change the values of "Security_options_edit" and "Security_zones_map_edit" to "0" to allow security options to be edited

3. Go to HKLM\SOFTWARE\Microsoft\Internet Explorer\Main\FeatureControl\FEATURE_RESTRICT_ABOUT_PROTOCOL

Change the value to 0.

Additional links

http://support.microsoft.com/kb/299838 -Unable to negotiate Kerberos authentication after upgrading to Internet Explorer 6
http://support.microsoft.com/kb/820780 - IE always prompt for authentication when browsing to web site already logged on to
http://msdn.microsoft.com/en-us/library/bb250462%28VS.85%29.aspx - Understsanding and working in Protected Mode in Internet Explorer

Internet Explorer Proxy Settings - Part 1

2009-11-11T11:19:00.000-08:00

inIf you frequently switch between office LAN, VPN and public internet connection, you may find that Internet Explorer occasionally gives you connectivity problem. If you are lucky enough, Internet Explorer's "Automatically detect settings" you do not have to bother with any proxy settings.

However, if that does not work then you have to manually set the proxy settings.

Sometimes I need to download files from the Internet when I remotely log in to Windows servers. By default Internet connectivity is disabled and there is no proxy settings. I find it a chore to set the proxy every time.

So here's a WSH vbscript you can use to automatically set the proxy settings for Internet Explorer:

1. Create a file with .vbs extension using your favorite text editor. Example: set_proxy.vbs
2. Copy these contents and save the file

' Change proxy-server.company.com:8080 to the correct proxy server name and port
' Change *.company.com; to all the intranet sites that will not go through the proxy

set WSHShell = WScript.CreateObject("WScript.Shell")

WSHShell.RegWrite "HKEY_CURRENT_USER\Software\Microsoft\Windows\CurrentVersion\Internet Settings\ProxyServer", "proxy-server.company.com:8080"
WSHShell.RegWrite "HKEY_CURRENT_USER\Software\Microsoft\Windows\CurrentVersion\Internet Settings\ProxyEnable", 1, "REG_DWORD"
WSHShell.RegWrite "HKEY_CURRENT_USER\Software\Microsoft\Windows\CurrentVersion\Internet Settings\ProxyOverride", "*.company.com;"

3. Execute the WSH vbscript file, and you should see this

If you click on "Advanced" you will see the list of sites which will not be accessed through the proxy.

For more information, check out these links:

http://www.pctools.com/guides/registry/detail/292/
http://technet.microsoft.com/en-us/library/cc736412%28WS.10%29.aspx

How to uninstall application from command line in Vista

2009-10-19T06:54:00.000-07:00

I was doing some application testing on a Vista laptop, and the "Programs and Features' window does not contain a single, installed appliaction. Not even logging in as Administrator and running "appwiz.cpl" solved the problem. If anybody knows the solution, I will really appreciate if you can share it with me.

Anyway, some Googling showed that it is possible to use wmic to uninstall the application.

1. Start wmic. In Windows command line, simply type "wmic"

2. List out all installed applications

wmic:root\cli>product get name

3. Once you have the exact application name, use this command:

wmic:root\cli>product where name="Adobe Reader 9" call uninstall

Location of Default.rdp file

2009-10-05T09:12:00.000-07:00

I frequently use Microsoft Terminal Services Client to connect to a couple of Windows servers, and find it annoying that the screen resolution always defaults to a larger size than my local display resolution.

I (stupidly) have the impression that the connection to each server is stored in its individual .rdp file, located somewhere in the Windows profile folder. After a long search and googling, I still cannot find any .rdp file.

It actually took me using SysInternal Process Monitor to find out mstsc.exe is writing to %My Documents%\Default.rdp !

So it is a simple matter of changing the values of resolution in the Display tab of the RDP client.

Additional information on Remote Desktop Connection Display - http://technet.microsoft.com/en-us/library/cc772472%28WS.10%29.aspx

Microsoft Paint tricks

2009-09-10T07:00:00.000-07:00

I use Microsoft Paint frequently to capture screen shots to send to my colleagues or to update documentation. I have a default "Temporary images" folder and I always prefer PNG for the simple reason of lossless compression. Here are some hacks that I come across as useful:

Set default "Save As" folder for Microsoft Paint

1. Start up Registry Editor
2. Go to HKEY_CURRENT_USER\Software\Microsoft\Windows\CurrentVersion\Explorer\Shell Folders\
3. Change the value of "My Pictures" to the folder you like

Set default "Save As" format to PNG

I have not found the solution yet. Does anyone knows?

SQL Server 2005 -> 2008 upgrade error - The installed product does not match the installation source(s)

2009-07-27T10:41:00.000-07:00

Saw an interesting article about a common error when upgrading SQL Server:

SQL Server Setup prompts with “The installed product does not match the installation source(s).”

Cross-OS filesystem on external USD Hard Disk

2009-05-12T22:59:00.000-07:00

Last Friday I went to Low Yat Plaza to buy an external USB hard disk. What motivated me was the realization that for the past 1 year I have been only focusing on work in office, and not learning much of anything that is not company-specific. I intend to experiment with newer OS, as well as having an external store to for my laptop - most of the space is used up by Vista.

As with most external disk drives, all 320 GB of it is formatted in FAT32. Which works fine if one uses it to store files and perform back up. I wanted to have 3 separate partition:

First partition (25%) to install a Windows OS
Second partition (25%) to install a *NIX OS - say Linux or FreeBSD
3rd Partition (50%) for storing files

Hence begun my journey to attempt to partition an external USB disk. First downloaded Knoppix, and used fdisk to create 3 partitions. I manage to create the 3 partitions, but Vista is unable to read the 2 newly created logical partitions. So I had to reformat it - which defaults to NTFS format. I thought it is simple enough to convert from NTFS back to FAT32 - I was wrong.

Next was using built in Windows utilities: Windows Vista's (Enterprise, SP1) Disk Management utility, which has rudimentary partition capabilities. Unfortuantely, it only allows 2 formats: NTFS and exFAT. I was looking for FAT32.

Then I tried the diskpart utility, which is bundled with Vista. Again, partitioning to NTFS works, but it will not allow me to partition to FAT32.

This is most probably due to Microsoft's push to newer file formats - NTFS and exFAT. I thought an older operating system will do the trick. I tried on Windows 2000 (Professional, SP4) using the "format" and "convert" utility. No such luck, it only a maximum size of 8 GB - probably using 512-Byte clusters and FAT16?

I was too fed up with all the wrangling and limitations of Windows tools. I will just live with NTFS, and when accessing the partitions from Linux will use the NTFS-3G driver to avoid corruptions during data writes.

Incorrect NDR message from Exchange 2007 server?

2009-05-05T10:17:00.000-07:00

Recently I was trying to email a user whose account was recently deleted. Instead of the usual error:

#550 5.1.1 RESOLVER.ADR.RecipNotFound; not found ##

I got a "Not Authorized" error instead"

#550 5.7.1 RESOLVER.RST.NotAuthorized; not authorized ##

Turned out the account was recently deleted, and my company has a policy of moving a deleted account into a "Obsolete User" AD container for 35 days before permanently purging the AD account.

The NDR message is definitely inaccurate.

Is this another of Microsoft's fault or lazy/mis-configuration from the Exchange folks?

Upgrading SQL Server 2005 Express to Enterprise edition

2009-05-05T01:52:00.000-07:00

I have been using SQL Server 2005 Express for a year now to test the back-end of an application, and now upgrading to the Enterprise edition to perform some ETL tasks in SSIS. So I will be trying to upgrade directly from Express edition to Enterprise without un-installing to see how it goes ;-)

Spefically, I am running SQL Server 2005 Express SP3, version 9.00.4035.00 . You can get the information by running this query:

SELECT SERVERPROPERTY('productversion'), SERVERPROPERTY ('productlevel'), SERVERPROPERTY ('edition')

Attempt 1: Direct installation of SQL Server 2005 Enterprise Edition

I simply run the set up file for Enterprise, resulting in this warning message

- Edition Change Check (Warning)
Messages
Edition Change Check
To change an existing instance of Microsoft SQL Server 2005 to a different edition of SQL Server 2005, you must run SQL Server 2005 Setup from the command prompt and include the SKUUPGRADE=1 parameter

I still can proceed with the installation, although I think I will run into a couple of issues. For example, I cannot select most of the server components:

OK, time to use SKUUPGRADE=1

Attempt 2: Running installer with SKUUPGRADE=1 option

Instead of double-clicking on the setup.exe file, I ran this from the command line

setup.exe SKUUPGRADE=1

A new warning message:

- Default Installation Path Permission Requirement (Warning)
Messages
Default Installation Path Permission Requirement
To run SQL Server Setup, log on to this computer with an account that has administrator privleges.

I am going to ignore that because I know I am logged on with an account with Adminsitrator rights.

After clicking Next, I still got the same issue as above. Server components greyed out.

Perhaps the problem is with the Enterprise edition?

Attempt 3: Using SQL Server 2005 Standard edition

Using this command line:

setup.exe SKUUPGRADE=1

Aha! Now I can see the server components:

When I look at the Operating Systems requirement in SQL Server Setup Help file, I found out that Enterprise Edition does not run on Windows 2000 Professional Edition SP4, which is what my laptop is running. Forgot rule #1: Always RTFM

Summary

Direct upgrade of SQL Server 2005 Express to SQL Server 2005 Standard works, but always check the setup requirements in the help files.

Difference between Local and LocalLow AppData folder ni Vista

2009-04-29T02:05:00.000-07:00

Short answer:

AppData\Roaming = files that follow user's login to all computers
AppData\Local = files specific to that computer
AppData\LocalLow = an area for low integrity apps to write to, e.g. Internet Explorer add-ons

Friendlier explanation: http://www.vistapcguy.net/?p=173

MSDN references:
The LocalLow folder may not be created on a Windows Vista SP1
Windows Integrity Mechanism Design
Designing Applications to Run at a Low Integrity Level

How to find out MS SQL Server version

2009-01-02T00:21:00.000-08:00

Run this query:

SELECT SERVERPROPERTY('productversion'), SERVERPROPERTY ('productlevel'), SERVERPROPERTY ('edition')

and you will get a result similar to this

Product Version Product Level Edition
9.00.3042.00 SP2 Express Edition

Another useful query:

SELECT @@VERSION
GO

Will show you this:

Microsoft SQL Server 2005 - 9.00.3042.00 (Intel X86) Feb 9 2007 22:47:07 Copyright (c) 1988-2005 Microsoft Corporation Express Edition on Windows NT 6.0 (Build 6001: Service Pack 1)

How to Remove Vista's Remote Desktop Connection warning

2008-10-01T09:01:00.000-07:00

I have been using Vista for the last few months, and one of the irritating symptoms I face is when Remote Desktop Connection (mstsc.exe) will prompt this warning when connecting to a Windows 2000 or Windows 2003 server:

Remote Desktop cannot verify the identity of the computer you want to connect to. This problem can occur if:

1) The remote computer is running a version of Windows that is earlier than Windows Vista
2) The remote computer is configured to support only the RDP security layer

The solution is to add a registry key at HKEY_LOCAL_MACHINE\SOFTWARE\Microsoft\Terminal Server Client

Type: DWORD (32-bit) Value
Value: 0

For more information, refer to this blog entry.

How to look up LDAP server

2008-04-07T01:18:00.000-07:00

In the global environment that I am supporting, I need an easy way to look up LDAP servers in a particular site. Learned this useful nslookup command from an engineer:

nslookup -type=all _ldap._tcp..europe.company.com

C:\Documents and Settings\hanxue.lee\Desktop>nslookup -type=all _ldap._tcp.EU-Ab
erdeen-ABE._sites.europe.company.com.
Server: dns-server.asia-pac.company.com
Address: 10.10.178.18

_ldap._tcp.EU-Aberdeen-ABE._sites.europe.company.com SRV service location:
priority = 0
weight = 100
port = 389
svr hostname = ldap-03003.europe.company.com
_ldap._tcp.EU-Aberdeen-ABE._sites.europe.company.com SRV service location:
priority = 0
weight = 100
port = 389
svr hostname = ldap-03001.europe.company.com
_ldap._tcp.EU-Aberdeen-ABE._sites.europe.company.com SRV service location:
priority = 0
weight = 100
port = 389
svr hostname = ldap-03002.europe.company.com
ldap-03003.europe.company.com internet address = 10.10.130.47
ldap-03001.europe.company.com internet address = 10.10.130.41
ldap-03002.europe.company.com internet address = 10.10.130.42

Regular expressions in grep, OR search

2008-03-18T19:52:00.000-07:00

I was trying to grep through the NIS group list to filter out group name by GID, and was scratching my head when this did not work:

ypcat group| grep -E "5520|110115|37650|33639"

Turn out that the version of grep in /usr/bin does not support regular expression, and the simple solution was:

ypcat group| /usr/xpg4/bin/grep -E "5520|110115|37650|33639"|awk -F: '{print $1}'

Argggh... n00b mistake

Kernel - 9 March 2008

2008-03-09T05:21:00.000-07:00

It must have been ages since I touched the kernel, and I mean for simple things like compilation and patching. Jeez.. it really hurts your technical skills if one stays in the Helpdesk for too long.

Anyway, I was playing around on an HP8200 box with RHEL 4 (customized) with this kernel:

Linux xxxxxx 2.6.9-55.0.2.ELsmp #1 SMP Tue Jun 12 17:58:20 EDT 2007 x86_64 x86_64 x86_64 GNU/Linux

Wanted to have a few virtual machine to mess with, but VMWare requires license, and somehow I fail to get xen to work. So its back to the original source from http://www.kernel.org Downloaded version 2.6.24, and it took this long to compile:

21m 14.649s

That's with 3 Intel(R) Xeon(TM) CPU 3.60GHz, and 6G of RAM. A bit slow?

I find the new Makefile for the kernel to be interesting, much cleaner, clearer output and best of all it makes rpm and deb packages too.

... back to preparing for LPI 102 and 201. (5 days left)

Findind and changing file ownership

2006-08-17T17:10:00.000-07:00

$ find /path -user < olduser> ¦ xargs chwon <newuser>

Flummoxed by IT

Ruby: Nokogiri installation problem when using Scrubyit

Related articles

Using RVM to control Ruby environment

Related articles

Node.js error ENOTFOUND, Domain name not found

Python and Ubuntu

How to manually install Django and pip

Related articles

How to install npm on Ubuntu

How to install Node.js on Ubuntu

Related articles

Set up Vim for Groovy development

Synchronizing PandaForm with Google Spreadsheets

Related articles

Formspring is asking for way too much information!

Related articles

Visibli causing duplicate Tweets

Related articles

Generating total number of requirements for each Requisite Pro projects

Workaround for Bitnami's custom PATH

Adventures installing google-api-client on Bitnami EC2 instance

How to sync Phone and SIM contacts with Google on Android phone

Cannot hibernate in Windows Vista

Eclipse Android error - An SDK Target must be specified

Vista poor performance after installing patches

tasksel error

IIS .NET compatibility error: Child nodes not allowed

yforth - is this a joke?

bash script to query AWS EC2 meta-data

Ignore GPG Key authentication when installing from Debian repository

Apache and Tomcat

Find out which process is using the most CPU

How to terminate processes from command line in Windows 2008

Manually start Language Bar in Windows Safe Mode; Putty behind firewall

Compacting Inbox / outlook.ost

Compacting PST

SQL Server Myths

Querying SQL Server backup results

Reset Terminal Service Client license

How to open .jar files as an archive folder?

Email Stubbing

Firefox and Windows NTLM/Kerberos authentication

Internet Explorer Proxy Settings - Part 2

Internet Explorer Proxy Settings - Part 1

How to uninstall application from command line in Vista

Location of Default.rdp file

Microsoft Paint tricks

SQL Server 2005 -> 2008 upgrade error - The installed product does not match the installation source(s)

Cross-OS filesystem on external USD Hard Disk

Incorrect NDR message from Exchange 2007 server?

Upgrading SQL Server 2005 Express to Enterprise edition

Difference between Local and LocalLow AppData folder ni Vista

How to find out MS SQL Server version

How to Remove Vista's Remote Desktop Connection warning

How to look up LDAP server

Regular expressions in grep, OR search

Kernel - 9 March 2008

Findind and changing file ownership