Who’s On Phirst

Cloud computing - one announcement and one rumor caught my attention this week

Amazon announced an Eclipse plug-in that helps you deploy Java applications to a Tomcat instance. The Eclipse plug-in takes advantage of the Eclipse WebTools project and marries it with a few custom views that show you EC2 (Elastic Compute Cloud)and EBS (Elastic Block Storage) specific information. The use case here is for Java developers to be able to deploy their Java applications to a Tomcat instance running on an EC2 instance in the cloud.

The second is a rumor that Google will have Java support for App Engine soon. Given Google's commitment to Eclipse already through GWT, I expect them to also have an Eclipse plug-in, as well.

As a Java engineer, I love this. For better or worse, Java is the lingua franca of business logic. And every engineer that speaks it does so with Eclipse. (My apologies to IntelliJ users.)

However, as a former SCM, I'm terrified of Java engineers providing me with a new application that runs on technology I have no experience with. If I were to put my IT hat on, I can imagine going to the mat with my VP of Development and not allowing my company to use ANY cloud resources. "We already have the servers." "This violates our security policy." "The SLA is more lax than our current SLA." The next thing you know, that Java app targeted for Tomcat in the cloud is now running on my existing Tomcat installation. Internal IT: 1, Cloud: 0.

It seems that both Amazon and Google are putting too much emphasis on targeting developers at the expense of IT operations. Personally, I think you must target all points of the application lifecycle to be successful. Take a look at what Microsoft does as an example; regardless of how you interact with Microsoft products (developer, IT, user), there are targeted tools just for you and your specific use cases.

I applaud both Amazon and Google's efforts. Their next offerings should be tools that allow you to migrate existing infrastructure to the cloud, not create new applications. They should help the IT operations folks move the web application running on a dusty beige box in the corner to the cloud. Instead of having IT migrate from off-lease hardware to new hardware, provide the tools to help them migrate to the cloud. Or, my personal favorite use case, mirror existing hardware to the cloud to act as a warm backup for disaster recovery.

At some point, those use cases will be implemented. The risk to Amazon and Google is that another player will provide the cloud AND the management tools out-of-the-box.

In Cloud Computing
Comment (0) Read More...

In Untagged 
Comment (0) Read More...

You really can’t beat Austin in the Spring. This week has made me especially happy to live here. We have tons going on right now – SXSW, Flatstock, Rodeo, Spring Break, March Madness, bluebonnets, a new awesome food truck on South Congress, Mighty Cone….this city is BUZZIN! Oh and did I mention that it is in the 80’s and sunny and that we can hear SXSW bands from our parking lot? Life is Good!

Phurnace had some great news this week. We were included in a Gartner report of Cool Vendors in IT Operations and Virtualization. Take a few seconds to check out the report. Once you are done reading that, you can check out the links above and book your next trip to Austin!

In Untagged 
Comment (0) Read More...

Back at the beginning of the company, we here at Phurnace made some pretty basic decisions that have shaped and defined what it is we do and how we go about doing it. At heart, Phurnace’s products get our customers out of the business of being middleware experts (WebSphere, WebLogic, Jboss, WebSphere Portal, Tomcat, etc.) and let’s them focus on THEIR apps, not the apps running their apps. We take care of that piece for them. And we are pretty good at it.

What we don’t do is a lot of the management around that. Single sign on. Tight ACL management. Automated heuristical compliance checks. Work flow, etc. Sorry, that’s not us. But there are a lot of great products out there that do those things, and that’s why we have - from day one - been fanatical about making Phurnace products easy to integrate.

And when talking about integration, there are really two things that are the most important: standards and flexibility. So our data model is pretty straightforward XML. We use log4j. And standard out. Basic stuff. And you can call us from batch, shell, ANT, or integrate directly with the Java API. Up to you. All the functionality is there for you regardless of your choice. Heck, we leverage that internally as well. Our Eclipse RCP UI calls our execution engines in the exact same way that an external ANT task would.

At Phurnace, just about every one of our customers wants us to integrate our products into their existing infrastructure. BMC BladeLogic, HP Opsware, IBM Rational BuildForge, or any one of a host of other products, or an internal build/deploy process. And we want them to. Because when you drop Phurnace into any one of those frameworks you get to do some very very cool things. For example, with Phurnace and BladeLogic together, you get the ability to set security at the object level in WebSphere, and have different users only have access to specific applications, not the “all or nothing” paradigm that current WebSphere customers have. And I think that’s pretty neat.

We keep our focus on what we are good at, and make it easy to integrate to products that other companies have built. And as long as we all get along, share, wash our hands, and be considerate of others then our customers get not only the value from our products, but increased value from the stuff they have already bought.

In Untagged 
Comment (0) Read More...

We had a discussion the other day with a very large bank and learned that they spend a huge amount of time and money on setting up test servers and fail-over servers, yet they have major problems assuring that the configurations match. In fact, they gave us an example of how their customer facing application dropped recently because their main site dropped and the fail over was not configured correctly. BAD. To keep this from happening again, they were planning to throw people at the problem. There is a better way.

It is now clear that Phurnace Deliver can also provide an “Agile Test Infrastructure” for higher test server utilization and dramatically lower costs.

The way it works is that large customers keep many different test environments to be responsive to test scenarios because it takes too long to configure a different environment required for a given test scenario that arises.

At the above mentioned bank, test server utilization is only 10%. To eliminate half of their test servers and the associated “hard dollar” hardware and software costs, Phurnace only needs to drive their utilization to 20%. The way Phurnace accomplishes this is by eliminating lesser used test servers and configuration environments and then by reconfiguring available servers to the required configuration on an as-needed basis.

Bottom line, Phurnace enables an “Agile Test Infrastructure” as well as enabling Agile Development.

In Agile Software Development
Comment (0) Read More...

I thought I would provide a step by step instruction on setting up IBM WebSphere Portal 6.0. This is how we do it here at Phurnace. I thought many of you could benefit from our example.

Create Linux Virtual Image

The first step is to create a virtual image to work with. For purposes of this example, we will create an image called "wp60source". This linux image will have a static ip.

Create the Image

cd /opt/Virtual\ Machines

• Make a directory for your virtual machine. We call ours wp60source in this example

mkdir wp60source

• For linux, copy the contents of the rhel4Base image to your VM's directory

cp -rf rhel4Base/* wp60source

• Login to phurnacedev03 with the VMWare Console
• Add the virtual machine to the list
  • Go to File->Open.
  • Select Browse...
  • Double click "wp60source".
  • Select the .vmx file contained within that folder

Configure the Virtual Machine Settings

• Edit Virtual Machine Settings
  • Bump Memory to 2G
  • Add 4 GB Virtual Disk
    • Click Add, Select virtual disk and accept defaults

Configure the Virtual Machine

• Startup the Virtual Machine
• When prompted about the Network Configuration, press any key
  • Remove Configuration
  • Add Configuration.
    These are my example settings , use your own IP address.

IP                   10.1.1.158
NetMask              255.255.255.0
Default Gateway      10.1.1.1
Nameserver           10.1.1.48

• Set the Hostname
  • Applications -> System Settings -> Network
  • Click DNS
  • Change Hostname
  • Reboot

• Extend the Logical Volume
  • Applications->System Settings->Logical Volume Management
  • Under "Unitialized Entities", Find the entity and initialize it. In this example, /dev/sdb1
  • Add this entity to the VolumeGroup00
  • Go to the Volume Group Properties and extend the slider.

• Bump up the ulimit ceiling for the number of open files
  • vi /etc/security/limits.conf
  • add the following lines

*    soft    nofile   10240
*    hard    nofile   10240

Extract the Installer for WebSphere Portal 6.0

Consult Disk Image Page for help regarding Disk Images.

• Find the installation images for Portal 6.0. First consult the

cd /mnt/install/WebSphere Portal/Portal_6.0
mkdir /opt/images
cp C93MXML.taz C93N3ML.taz C93LSML.zip C93M4ML.zip  C93LUML.zip  C12YDML.zip /opt/images

• Extract each of these images in their own folder

cd /opt/images
mkdir C93MXML
cd C93MXML
unzip ../C93MXML.zip

mkdir C93MXML
cd C93MXML
tar zxvf ../C93MXML.taz
...

• Rename each of the diretories to the names provided on the Disk Image Page.
  This also verifies you grabbed the correct zips.

mv C12YDML W-Setup
mv C93MXML IL-1
mv C93N3ML IL-2
mv C93LSML IL-3
mv C93M4ML IL-4
mv C93LUML IL-5

Install Portal

In the W-Setup directory, run the installer.

./install.sh.

Choose Typical.
Provide a name of the Cell, Node,
Provide a WAS user/pass
Provide a Portal Admin user/pass
Accept defaults.

Install Patches

• You may need to free up space and delete the disc images from the base install

• Copy over the patch installer and fix paks. These are contained on the install drive under WebSphere Portal in the 6.0.1.3 update directory.

cd /mnt/install/WebSphere\ Portal/
cp -rf 6.0.1.3\ update/ /opt/images

• Extract the "WebSphere Update Installer" Installer

mkdir WasUpdateInstaller
cd WasUpdateInstaller
unzip ../6.1.0-WS-UPDI-LinuxIA32-FP0000017.zip

• Run the "WebSphere Update Installer" Installer and install to /opt/IBM/WebSphere/AppServer

• Copy the fixpacks into the maintenance directory

cd /opt/images/6.0.1.3 update
cp *.pak /opt/IBM/WebSphere/AppServer/updateinstaller/maintenance/

• Now run the WebSphere Update Installer

cd /opt/IBM/WebSphere/AppServer/updateinstaller
./update
Relaunch for each of the .pak files

• Unzip the "Portal Update Installer" Installer

• Move the directory to /opt/IBM/WebSphere/PortalServer/updateinstaller
• Unzip the fix pak.
• Move the contained files to /opt/IBM/WebSphere/PortalServer/updateinstaller/fixpaks
• Update the passwords in /opt/IBM/WebSphere/PortalServer/config/wpsconfig.properties
• Run . ./setupCmdLine.sh from /opt/IBM/WebSphere/AppServer/bin
• Run the Portal Update Install Wizard
• Select the jar file in the fixpaks directory

In WebSphere Portal
Comment (2) Read More...

Seems that there’s a lot of talk in the blogsphere and around the water cooler now a days about application virtualization being the next step in the abstraction revolution taking over the data center. Whether to current infrastructure, virtual images, or now out to the cloud, companies are really starting to wrap their heads around the important differences between the data center centric perspective of days past and the application centric view point inching into strategic conversations.

So, what is Application Virtualization? Realistically, it’s not that much different from virtualization going on right now, just at a different layer of the stack. Instead of being “bare iron and up” it’s about making smaller, more surgical changes to applications as new versions come out. It’s about parsing apart the application (and what it needs in terms of pre-reqs) from the underlying environment that it is going to run on. In fact, I would say that there is really nothing different between app virtualization than infrastructure virtualization – except the target layer.

Why is that important? Well, the environment has become a lot more nebulous. In a way application virtualization is a response to infrastructure virtualization. Where the app is going to land is now a lot more uncertain than in days past. Applications have had to become more flexible as the environment has become more amorphous. And, as I suspect will happen, the cloud becomes a bigger component of companies larger IT strategy, application abstraction and virtualization will become a necessity, and not just a productivity booster.

Companies like BMC (BladeLogic), IBM (Tivoli), HP (Server Automation - Opsware) are starting to take notice of this shift. Operations tools can’t just be about laying down the OS and getting the network configured. The churn is inside applications now, and that’s where the incremental value of such tools really lies. I would look to see major initiatives from all of them around application virtualization soon.

In Virtualization
Comment (0) Read More...