Monday, March 31, 2014

Using passwd command to lock and unlock Linux Account

To lock a linux account, you can use the command

# passwd -l username
Locking password for user username.
passwd: Success

You will notice on the /etc/shadow file, that there will be a ! before the encrypted password string

To unlock the linux account, you can use the command
# passwd -u username
Unlocking password for user username.
passwd: Success.

Sunday, March 30, 2014

Tectonic shift as public cloud giants acknowledge the power of private deployment options

I thought the article Tectonic shift as public cloud giants acknowledge the power of private deployment options from GIGAOM is quite significance as the article mention that the

Both Amazon and Google, which launched huge price cuts and a raft of new services on Tuesday, clearly have the scale and resources to vacuum up a lot more corporate jobs into their respective clouds. What both lack is a full hybrid cloud strategy that would let businesses balance private and public cloud deployment as they see fit. And that’s public-private cloud coexistence message is the one legacy IT players like VMware and Microsoft have been pounding home.

But the public cloud companies are starting to get there. Google announced much faster data ingestion for its BigQuery database service that can now take in 100,000  rows per second versus 1,000 before.
The ability to pump data into and out of a public cloud whether from another public cloud or a private cloud or an on-premise database is a huge deal for enterprise accounts and something they want to be easy so they can avoid cloud lock-in.

Other related posts:
  1. A peek into NTU's supercomputer and hybrid cloud  
  2. 50% Enterprise to Use Hybrid Cloud by 2017

Friday, March 28, 2014

Shipments of Disk Drives Declines in 5% in 2013

I refer to this article Shipments of Disk Drives Decline 5% in 2013 from . According to the article,

A near-doubling of the market for SSDs in 2013 was unable to help the computer-related storage industry last year, which slid 5% in shipments because of continuing contractions in the hard-disk drive and optical disk drive segments, according to a new report from IHS Technology.

(Source: IHS Technology, January 2014)'

HDD down but not out
In the HDD segment, the rise of smartphones and tablets has dented the once-powerful appeal of computers, impacting HDD volume. The losses are especially apparent in the so-called client PC market-the consumer side of the PC business.

Wednesday, March 26, 2014

Kipmi0 May Show Increased CPU Utilization on Linux

This article is takenf from IBM Support Portal. Kipmi0 May Show Increased CPU Utilization on Linux


The kipmi0 process may show increased CPU utilization in Linux. The utilization may increase up to 100% when the IPMI (Intelligent Platform Management Interface) device, such as a BMC (Baseboard Management Controller) or IMM (Integrated Management Controller) is busy or non-responsive.


No fix required. You should ignore increased CPU utilization as it has no impact on actual system performance.


1. If using an IPMI device, reset the BMC or reboot the system.

2. If not using an IPMI device, stop the IMPI service by issuing the following command:

    service ipmi stop


IBM has requested that the design of kipmi be reviewed in future releases. Any changes that may occur through this review would be implemented in future Linux Distribution updates.

The increased CPU usage of kipmi is normal. The hardware interface is not interrupt driven, so the driver must poll the device for status and messages. It is this polling that is showing up as a busy CPU. The kipmi kernel thread has its priority intentionally lowered so that it does not interfere with other processes on the system. Even when it is polling in its tightest loop (usually when it thinks the BMC has active events it needs to handle), it will give up the CPU to any process that wants it.

The CPU usage seen for the kipmi0 kernel thread is what is known as IDLE time. Kipmi0 is running when there are no other jobs to run and is one of the lowest priority processes on the system.

This is the standard behavior for this interface and is not unique to this process running in an IBM environment or specifically with IBM Systems Director.

Tuesday, March 25, 2014

Configuring vSphere Flash Read Cache

This article Configuring vSphere Flash Read Cache is taken from It is definitely a good read. One of good features of VSphere 5.5 is the Flash Read Cache.

According to the article,

This capability allows a VMDK (the VMware virtual disk format) to have a "slice" of higher performing SSD storage without putting the entire contents on the SSD. This only applies to reads, but it can be a great way to get a big boost in I/O that is shared across multiple VMs.

The picture summarise it so well

Do read the article for an in-depth explanation. 

Saturday, March 22, 2014

Intel Xeon Phi Code Recipes, workshops and tutorials

Do take a look at Intel Xeon Phi Coprocessor Site where there are code recipes, technical case studies, video workshop and tutorials. This is very useful for those interested in using Intel Xeon Phi Coprocessor for computing

Friday, March 21, 2014

Best Ways to Learn Xeon Phi Programming

For those who are interested to learn Xeon Phi Programming, you may want to take a look at Best Ways to Learn Xeon Phi Programming

Thursday, March 20, 2014

Subnet Calculator

Subnet Calculator from is a good online tool to calculate subnet

Wednesday, March 19, 2014

Summary of Job Management Commands for MAUI

This is a good summary taken from Adaptive Computing 4.3 Job Managament Commands to manage jobs for MAUI.

Command Flags Description
cancel existing job
display job state, resource requirements, environment, constraints, credentials, history, allocated resources, and resource utilization
diagnose -j display summarized job information and any unexpected state
releasehold [-a] remove job holds or defers
start job immediately if possible
set hold on job
set/modify QoS of existing job
adjust job/system priority of job

Wednesday, March 12, 2014

Hot Water Cooling Technologies

This article is quite interesting and you may want to read it "Water Cooling Comes of Age, Again" by Larry Vertal

Here are some interesting excerpts

Many “water cooled” data centers are using rear door or in-row coolers. While rear door and in-row coolers are a step in improving cooling efficiency they provide only intermediate benefits as they still require cold water.  These systems in effect bring the CRAH unit closer to the server, reducing the energy needed to move air through the data center.  The servers are still air cooled.

Bringing liquid inside the server to directly cool high heat flux components such as CPUs, GPUs and memory provides even greater benefits as is done with Asetek’s Direct-to-Chip (D2C) technology.  These benefits include:
  • Enabling much higher rack densities.
  • Reducing overhead power requirements of the data center for cooling.
  • Thereby reducing operational costs and green house emissions.
  • Lowering acoustic noise.

Hot Water Cooling is Energy Efficient and EffectiveThe term “hot water cooling” appears to be an oxymoron, but this method is one of the most effective. Consider that in order to operate properly the surface temperature of a CPU (case temp) needs to be maintained in the 67°C to 85°C (153°F to 185°F) range, depending on CPU model.  The operating surface temperatures for memory chips, GPUs and co-processors is even higher, in the 90°to 95°C (194°F to 203°F) range.  Cooling with air requires a high initial temperature difference between the cooling air and the desired surface temperature (high Delta T).  The cooling efficiency of water allows it to maintain the required case temps with a much smaller delta T.  This means the water used for cooling the components can be hot.
The use of hot water to directly cool the high heat flux components within servers such as CPUs, GPUs and memory is the foundation of an effective Direct-to-Chip, liquid cooling system. This technology replaces air heat sinks and puts cold plates directly on the hot spots inside the server.  Liquid is flowed over the cold plates, collects the heat and is pumped out of the server into the CDU (Cooling Distribution Unit).  At the CDU, heat is exchanged with facilities liquid and pumped out of the data center.

Cooling with hot water means that, even in summer, outdoor air is cold enough to cool the very hot return water from the data center back down to the temperature need to cool these components. Using outdoor or ambient temperature air to ultimately cool the data center is the basis of "free cooling." Outdoor air cools the water, generally by using dry coolers or cooling towers, meaning no power is used to actively chill the water. Additionally, the water leaving the data center is hot enough to enable waste heat recycling, further increasing energy cost savings and resulting in cooling Energy Reuse Efficiencies (EREs) below 1.

Monday, March 10, 2014

Checking version of the Intel® IPP, Intel® MKL and Intel® TBB Libraries included in the Intel® Composer Bundles (Updated)

This blog is an update of Checking version of the Intel® IPP, Intel® MKL and Intel® TBB Libraries included in the Intel® Composer Bundles

Do look at the updated link from Intel Compiler and Composer Update version numbers to compiler version number mapping

Composer XE 2013 SP1 Intel Registration
Activation Date
Windows Version / build Linux Version / build
Composer XE 2013 SP1 Update 2 2014.02.12 / 20140130 / 20140120
Composer XE 2013 SP1 Update 1 2013.10.18 / 20131008 / 20131008
Composer XE 2013 SP1 initial reiease 2013.09.04 / 20130728 / 20130728

Sunday, March 9, 2014

Uninstalling Torque

Let's say you made a wrong installation and you wish to do a fresh install of torque, you can to the installation directory and issue the command

# make uninstall

Later you can proceed to remove the directory
# rm -Rv --force /var/spool/torque
# rm -Rv --force /opt/torque
# rm --force /etc/init.d/ trqauthd
# rm --force /etc/init.d/pbs_server
# rm --force /etc/init.d/pbs_sched
# rm --force /etc/init.d/pbs_mom 

Friday, March 7, 2014

Compiler Support for the Fortran 2003 standard

If you need to check on the status of compiler support for the Fortran 2003 standard, do take a look at Fortran Wiki Compiler Support for the Fortran 2003 Standard

Fo great interest for me is of course Intel and GNU which are my preferred compilers. In the diagram, GNU 4.7/4.8 and Intel 14 (Intel 2013 SP1) is listed.

PGI 13.7 and IBM 14.1 is rather impressive with almost full compliance except for support for international character sets.

Thursday, March 6, 2014

PBS_MOM Error Mismatching protocols. Expected protocol 4 but read reply for 0

I encountered this error on my compute nodes using Torque 4.2.5.
pbs_mom.29384;Svr;pbs_mom;LOG_ERROR::read_tcp_reply, Mismatching protocols. Expected protocol 4 but read reply for 0
 This error was quite misleading. I was looking at my protocol which was IB and Ethernet.

 When I did a pbsnodes -l, all the compute nodes were down.
# pbsnodes -l
node-c00 down
node-c01 down

After some troubleshooting, I realised that the error is due to use of inconsistent use of short hostname and long hostname. On my /etc/hosts, I used the long hostname for the compute node first (which Torque Server pick up.    node-c00

But on each of the client nodes ie /etc/sysconfig/network, I used the short hostname. This create some confusion for the torque server


To correct the matter, just rename the HOSTNAME to the long name

Do a restart of the pbs_mom on the client node and you should get your nodes alive
# service pbs_mom restart

Monday, March 3, 2014

Using Ctrl-R to search for terminal history

Press Ctrl and R together and the linux terminal will give  (reverse-i-search)`': Type a search word inside and the command will try to find the most recent command match

For example, If I input 'ssh' (without the quotation)
(reverse-i-search)`':`ssh': ssh node-c00

For more information
  1. Ctrl+R to search and other terminal history tricks