Data Center Temperature: ASHRAE's Perspective

A interesting article from Data Centre Posts "Data Center Temperature: ASHRAE's Perspective"

On the Moisture Lower Limit
.....A lower limit of 5.5°C (41.9°F) was chosen which corresponded to the lowest absolute moisture level of the 2004 recommended limits.....

On the Moisture Upper Limit 
.....It was found that extended periods above 60% can result in failures. As a result, a relative humidity upper limit of 60% was chosen.....

Installing check_mk for Nagios on CentOS 5.x

check_mk is a wonderful "a new general purpose Nagios-plugin for retrieving data". But this wonder plugins is a good replacement for NRPE, NSClients++ etc. I've tried using check_mk in place of NSClient++ to monitor my Windows Machines successfully


Installing Nagios is straightforward. You may want to see Blog Entry Using Nagios 2.x/3.x on CentOS. In a nutshell, do this in sequence to avoid dependency issues

# yum install nagios nagios-devel

# yum install nagios-plugins-all

Downloading and unpacking check_mk

# wget http://mathias-kettner.de/download/check_mk-1.1.8.tar.gz

# tar -zxvf check_mk-1.1.8.tar.gz

# cd check_mk-1.1.8

# ./setup.sh --yes

Restart the Service

# service nagios restart
# service apache restart

Making the agent accessible through xinetd

# cp -p /usr/share/check_mk/agents/check_mk_agent.linux /usr/bin/check_mk_agent
# cp -p /usr/share/check_mk/agents/xinetd.conf /etc/xinetd.d/check_mk

Restart xinetd service.

# service xinetd restart

Centrify Express Express Tips and Tricks

Centrify Express provides a couple of good links to provide some quick-start to configure Linux Box to talk to the MS AD

1. For Best Practise Video, see Centrify Express Tips and Tricks
2. For the  community site including forum see Centrify Express Community Site

Testing the Infiniband Interconnect Performance with Intel MPI Benchmark

This writeup focuses on verifying the performance of the Infiniband Interconnects or RDMA/iWARP Interconnects with Intel MPI Benchmark. For more information, do look at my Linux Cluster Blog

1. Testing the Infiniband Interconnect Performance with Intel MPI Benchmark (Part I)
2. Testing the Infiniband Interconnect Performance with Intel MPI Benchmark (Part II) 

Enabling CGI on Apache 2.x

For more information, do ready up the Apache Tutorial: Dynamic Content with CGI which provides with a easy-to-read configuration if you wish to enable CGI on the Apache 2.x


The ScriptAlias directive tells Apache that a particular directory is set aside for CGI programs. Apache will assume that every file in this directory is a CGI program, and will attempt to execute it, when that particular resource is requested by a client.

ScriptAlias /cgi-bin/ /usr/local/apache2/cgi-bin/

Explicitly using Options to permit CGI execution

You could explicitly use the Options directive, inside your main server configuration file, to specify that CGI execution was permitted in a particular directory:

<Directory /usr/local/apache2/htdocs/somedir>
Options +ExecCGI
AddHandler cgi-script .cgi .pl
</Directory>

User Directories
To allow CGI program execution for any file ending in .cgi in users' directories with a cgi-bin extension, you can do the following:

<Directory /home/*/public_html/cg-bin>
Options +ExecCGI
AddHandler cgi-script .cgi .pl
</Directory>

Python Error "Could not find platform independent libraries"

I encountered this error after I configure Python.

Could not find platform independent libraries 
Could not find platform dependent libraries 
Consider setting $PYTHONHOME to [:]
ImportError: No module named site

Actually, this error can be eliminated once I did a make install. Sometimes, even we system admin forget the simplest thing to do. Sigh

After you untar the python package, do a

# ./configure
# make 
# make install

Resolution for bash orted command not found

If you faced an issue like this below, it is likely due to your path-ing issue in an non-interactive mode. See Blog Entry Unable to source path during non-interactive ssh session

bash: orted: command not found
--------------------------------------------------------------------------
A daemon (pid 12639) died unexpectedly with status 127 while attempting
to launch so we are aborting.


There may be more information reported by the environment (see above).


This may be because the daemon was unable to find all the needed shared
libraries on the remote node. You may set your LD_LIBRARY_PATH to have the
location of the shared libraries on the remote nodes and this will
automatically be forwarded to the remote nodes.
--------------------------------------------------------------------------

You can verify by using this niffy commands to check

$  ssh remote-server-name which orted

You should get something like the return value below if you have source correctly in your .bashrc and .bash_profile

$ /usr/local/mpi/intel/bin/orted

If not, you will get

which: no orted in (/usr/local/bin:/bin:/usr/bin)

Either you put your binaries in the standard path like /usr/local/bin /bin /usr/bin or you have to source correctly in your .bashrc and .bash_profile

Unable to source path during non-interactive ssh session

In a cluster, non-interactive batch mode, ssh session somehow do not "source" the PATH at /etc/profile or /etc/profile.d/*.sh when launched in an non-interactive mode. For non-interactive mode, ssh session only "source" path the standard binary path such as /usr/local/bin, /usr/bin etc

If your binaries which are placed in non-standard paths and are required to be sourced during the non-interactive session, remember to create .bashrc and .bash_profile file at your home folder and place your sourcing there.
For example:
source PATH=/usr/mpi/gcc/bin:$PATH

In an interactive mode, you should be able to source the PATH provided at /etc/profile or /etc/profile.d

Linux Mint 10 and Huawei Mobile Broadband E153

I have installed Linux Mint 10 on my HP 2140 MiniNote, which feature following upstream components: Ubuntu 10.10, Linux 2.6.35, Gnome 2.32, Xorg 7.5.


Working with Huawei Mobile Broadband E153 was a breeze. It detected the broadband very easily.

Repeating previous command using sudo

If you need to repeat the previous command with the sudo command, you may wish to use the command

#sudo !!

Save you a couple of keystroke :)

Gmond dead but subsys locked for ganglia monitoring daemon

Following the tutorial instructions from Installing and configuring Ganglia on CentOS 5.x ,I've installed ganglia and ganglia-gmond. However when I attempt to start the gmond service

# service gmond start

Somehow the service refused to start. When I try another way to start the service

# /etc/init.d/gmond start

I notice the error, and the service do not start

gmond dead but subsys locked

To Solve the Issue
Step 1: Turn the Debug Level at /etc/gmond.conf to 100 found at the "globals" section

globals {
daemonize = yes
setuid = yes
user = ganglia
debug_level = 100
max_udp_msg_len = 1472
mute = no
deaf = no
host_dmax = 0 /*secs */
cleanup_threshold = 300 /*secs */
gexec = no
}

Step 2: Start the Service again.We should be able to see the error that caused "gmond dead but subsys locked". In my case, it is due to somehow my passwd file does not contain ganglia user. After fixing this missing user group, I was able to restart the ganglia-gmond daemon sucessfully.

For additional Information  on /etc/gmond.conf, you may want to use man pages or go to http://linux.die.net/man/5/gmond.conf

Killing process from Top Utilities

You can use "Top" utilities to kill processes instead of killing from the command line. Just press "k" and the option "PID to kill:" appear. Just type the PID and press "Enter" on the keyboard. The option kill PID ...... with signal [15]: appears. Press "Enter" again and the Process will be eliminated.

Sorting Processes by using Top

Top is one of my favourite commands for the system administration. The most important columns I often look out for is PID, USER, %CPU, %MEM, NI

If you wish to do sorting on Top, you can use a short-key "Shift+F"

Select the column you wish to sort by......

Cool right?