Feedburner to be acquired by Google

May 24, 2007


I just heard that 'Feedburner' is going to be acquired by Google for around $ 100 Million.

Here is the Original Story...

Boot Windows XP from a USB flash driveBoot Windows XP from a USB flash drive

May 1, 2007

You can't boot Windows XP...


Takeaway: You can't boot Windows XP from a floppy disk the way you used to be able to with DOS. One handy way to easily boot XP is by using a USB flash drive. Here's how to make it work.

Almost everyone who has worked with computers for any length of time at all has run into at least one situation in which a problem left a PC unbootable. What if you could return the machine to a bootable state just by inserting a USB flash drive though? Believe it or not, it is actually possible to install a bootable copy of Windows XP onto a flash drive and then boot a PC off of the flash drive. From there, you can use applications that you have installed on the flash drive (anti virus, anti spyware, disk repair, etc.) to fix the PC's problem. In this article, I will show you how.

What's the catch?

As with most cool new techniques, there are a few catches. For starters, not every PC is capable of booting from a USB flash drive. For the most part, computers manufactured within the last two years are generally able to boot from a flash drive. Older systems may require a BIOS update, or might not be able to boot from a flash drive at all.

Another catch is that not every flash drive will get the job done. The primary factors that limit your use of a particular flash drive are capacity and speed. Technically, speed isn't really a limiting factor, but booting Windows will be painfully slow unless you use a flash drive that supports USB 2.0.

The flash drive's capacity is actually a limiting factor though. Surprisingly though, there are size limits on both the upper and lower end. Your flash drive can't be too large or too small. There isn't really a documented minimal size for a flash drive. You just need something large enough to hold Windows XP and a few applications. As you probably know, Windows XP normally consumes over a gigabyte of disk space. Later I will show you how to use a free utility to trim the excess fat off of Windows XP and make it a whole lot smaller. Even so, I still recommend that your flash drive be at least a minimum of 256 MB in size.

As I mentioned, there is a maximum size for the USB flash drive that you can use. Currently, USB flash drives exist in sizes of up to 4 GB, and 8 GB flash drives are expected to be available by the end of the year. As nice as it would be to have 8 GB to play with, the flash drive that you use for this project can be no larger than 2 GB. The reason for this is because you will have to format the flash drive using the FAT-16 file system, which has a 2 GB limit. Presently, you are stuck using FAT-16 because most computers will not recognize a flash drive as being bootable if the drive is formatted with anything other than FAT-16.

Preparing your Windows installation CD

One of the requirements for creating our bootable USB flash drive is a Windows XP with Service Pack 2 installation CD. If your Windows XP installation CD doesn't already include Service Pack 2, then you will have to make a CD that includes Service Pack 2 through a technique called slipstreaming.

Other requirements

In addition to your Windows XP installation CD, there are a couple of other things that you are going to need. For starters, you will need the HP USB Disk Storage Format Tool. You can download this tool for free.

Another utility that you are going to need is Bart's Preinstalled Environment Bootable Live Windows CD / DVD, or BartPE for short. You can download this utility for free from the BartPE Web site.

In addition to the software requirements, you must verify that the PC that you will be using to create the Windows deployment has 1.5 GB of free hard disk space (minimum) and supports booting from a USB device. I also strongly recommend that the PC be running Windows XP Service Pack 2. Prior to Service Pack 2, Windows XP sometimes had trouble interacting with USB storage devices.

Formatting the flash drive

Now that you have all of the prerequisites taken care of, it's time to actually start setting up our flash drive. The first step in doing so, as strange as it sounds, is to format the flash drive. Windows will actually let you format a flash drive in the same way that you format a floppy disk. However, formatting a flash drive in this way will not work for this project. Furthermore, using Windows to format a flash drive directly has been known to destroy some types of flash drives.

Instead, you must format the flash drive by using the HP USB Disk Storage Format Tool that you downloaded earlier. To do so, simply open the utility, select the device followed by the FAT file system option and click Start.

Once the device has been formatted, you must make it bootable. To do so, you must copy the BOOT.INI, NTLDR, and NTDETECT from the root directory of your PC's boot drive to the flash drive. These files are hidden by default, so you will either have to configure Windows Explorer to show hidden files (including protected operating system files) or you will have to open a Command Prompt window and use the COPY command to copy the files.

If you choose to use the Windows Explorer method, then open Internet Explorer and enter C: into the address bar so that you are looking at your local hard drive. Next, select the Folder Options command from the Tools menu. When the Folder Options properties sheet opens, select the View tab. Now, just select the Show Hidden Files and Folders and deselect the Hide Extensions for Known File Types and the Hide Protected Operating System Files check boxes. Click OK to continue.

Booting from the USB flash drive

Now that you have formatted your USB flash drive and installed the boot files onto it, the next thing that you must do is to configure your PC to allow you to boot from the flash drive. This is all done through the computer's BIOS Setup. I can't give you specific instructions for this part, because every computer is different. I can give you a few pointers though.

You can access your computer's BIOS by pressing a specific key immediately after you turn the PC on. The key varies, but it is usually either [F1], [F2], or [Delete]. Once you are in the BIOS Setup, you should verify that all of your computer's USB options are enabled. This might include things like support for legacy USB devices or support for USB 2.0. If there is a time out setting for USB devices, you should set it to the max to insure that the system doesn't time out while waiting on the USB device to boot.

Next, find the section on boot device priority. Normally, a USB flash drive (which is usually listed as USB-HDD, but may be listed as a removable device) will have a very low boot priority. If the USB flash drive's boot priority is lower than the hard disk (listed as HDD) then the only time the computer would ever boot off of the USB flash drive is if the system were to fail to boot from the hard disk. You must therefore rearrange the boot device priority so that the flash drive has a higher priority than the hard drive.

Configuring Windows

Now that we have finally made it through all of the prep work, it's time to start setting up Windows. As you have probably already guessed, the process of installing Windows to a flash drive is quite a bit different from your normal, run of the mill installation. There are a couple of reasons for this.

For starters, a full blown Windows XP deployment takes up over a Gigabyte of hard disk space. When you are installing to a flash drive, disk space is a scarce commodity. Even if you have over a Gigabyte of space on your flash drive, you probably don't want to use it all on Windows. It would be nice to have room to install a few applications. Therefore, you need to trim the excess fat off of Windows.

The other reason why the installation process is so different from the usual Windows installation is because Windows Setup is not designed to install Windows to a flash drive. You therefore have to configure Windows using an alternate method.

The PEBuilder utility that you downloaded earlier can take care of both of these issues. PEBuilder is designed to create a build of Windows XP (or Windows Server 2003) that does not take up as much space as a full blown installation. Once you create this new build, you can copy it to the flash drive. For right now, I will show you how to create a basic Windows build and copy it to the memory stick. Unfortunately, it's rather difficult to install applications once Windows is up and running. Therefore, after I show you how to create a basic Windows build, I will show you how to create a build that includes some applications.

Begin the process by opening PEBuilder. When you open PEBuilder, you will see a screen similar to the one that's shown in Figure A. Simply enter the path to the Windows installation files (the ones from your Windows XP with Service Pack 2 installation CD). Next, verify that the Create ISO Image and the Burn to CD check boxes are not selected and then click the Build button. PEBuilder will now create the new Windows build.

Figure A
Free Image Hosting at www.ImageShack.us
You must use PEBuilder to create a Windows build that will work with a flash drive.

Now, it's time to copy Windows to the flash drive. To do so, you will have to use a special batch file that's included with PEBuilder. Open a Command Prompt window and navigate to c:\pebuilder313\plugin\peinst. Now, insert an empty flash drive into the computer's USB port and then execute the file PEINST.CMD. You will now see a menu appear as shown in Figure B.

Figure B
Free Image Hosting at www.ImageShack.us
PEBuilder uses a batch file to install Windows onto a flash drive.

Type 1 and press [Enter] and you will be prompted to enter the path to the build that you have created. Enter C:\pebuilder313\BartPE. Now, type 2, press [Enter], and you will be prompted for the target path. Enter the drive letter that Windows has assigned to your USB flash drive. After doing so, the menu is updated as shown in Figure C. The menu now displays the source path and the destination drive. Type 5 and press [Enter] to install Windows to the flash drive.

Figure C
Free Image Hosting at www.ImageShack.us
Use menu option 5 to install Windows to the flash drive.

Installing applications

Now that I have shown you how to create and install a basic Windows build, I want to talk for a moment about how you can add an application to the build (prior to creating it). The PEBuilder program comes pre-configured to support a number of common Windows applications, but does not come with the applications themselves.

The reason why installing applications can be a little bit tricky is because most Windows applications modify the Windows registry. The build that you are creating is basically a collection of installation files, and the build itself does not contain a registry (the registry gets created when Windows is installed onto the flash drive). As such, PEBuilder uses a sort of registry emulator.

If you go to the C:\PEBUILDER313\PLUGIN folder, you will see sub folders for a number of different applications. If you open one of these application folders, you will see that the folder contains an INF file and a FILES folder. The INF file contains all of the information that would normally go into the registry, and the FILES folder stores all of the program's files.

To see how this works, let's install an application that I'm sure most of you are familiar with; Nero. Begin by installing Nero onto the machine that's running PEBuilder, as if you planned to run Nero locally on that machine. When the installation completes, copy all of the files from C:\Program Files\ahead\Nero to C:\pebuilder313\plugin\nero burning rom\files. In this particular case, the nero burning rom folder is the folder that has been set aside for the Nero application. The Files sub folder is intended to store Nero's system files.

Now, you must take care of Nero's registry entries. To do so, go to the C:\pebuilder313\plugin\nero burning rom folder and open the PENERO.INF file using Notepad. As I explained earlier, the INF file in an application's folder is used to store the application's registry entries. For Nero and all of the other applications that PEBuilder predefines, the INF file is pre-configured. You just have to make a few changes that are specific to your system.

In this particular case, the PENERO.INF file is designed to support both Nero versions 5.x and 6.x. Initially, the lines for both versions are commented out. You must therefore determine which version you have and then remove the semi colon from the beginning of the lines that apply to that version. If you look at Figure D, you can see how the two versions are separated.

Figure D
Free Image Hosting at www.ImageShack.us
An application's registry entries are stored in an INF file.

Once you uncomment the appropriate lines, just replace "Your Name", "Your Company Name" and "Your Serial Number" with your name, your company's name, and your Nero product key. Save the file, and your set to go. The next time that you click the Build button, Nero will be included in the build.

Putting XP in your pocket

Running Windows from a flash drive isn't an exact science. Sometimes the process just doesn't work and there is no good reason why. As more PCs start to support booting from USB devices though, USB boots should become more standardized, and the technique should become more reliable.




Red Hat author dishes on RHEL 5

Tammy Fox has worn many hats...


By Jack Loftus, News Writer
23 Apr 2007 | SearchEnterpriseLinux.com

Tammy Fox has worn many hats in the past for Red Hat Inc., including technical writer, lead of the documents team and founding editor of Red Hat Magazine. She has also been a professional Linux writer for the past seven years. In other words, she knows Red Hat Linux.

Fox's latest work, Red Hat Enterprise Linux 5 Administration Unleashed, arrived this month intended as a tome of knowledge for intermediate to advanced Linux system administrators. Before it was published, Fox sat down with SearchEnterpriseLinux.com to provide some insight on the latest version of the Linux community's leading commercial distribution, Red Hat Enterprise Linux 5 (RHEL 5).

What general best practices or tips can you share to make an upgrade to RHEL 5 a smooth one?

Tammy Fox: Performing an upgrade with the installation program basically updates any existing RPM packages on your system. If the new version of the package includes changes that make your existing configuration invalid, some packages will try to migrate your existing data, but it is not required.

The best practice is to have your data on a dedicated storage system that does not get reformatted during a fresh installation. Depending on the size of your organization, this can be a SAN, a data cluster, or just a set of dedicated hard drives in the system. Back up any existing data and configuration files, and perform a fresh installation. Restore your configuration files one by one, testing each program. Be sure to pay close attention to whether or not the software version has changed from your previous version of RHEL to RHEL 5. Finally, reconnect your data storage solution and perform any additional tests necessary.

Of course, all of these steps should be performed on a test system before the plan is implemented on a production system.

What are some key differences between what administrators see in RHEL 4 deployments, and RHEL 5 ones?

Fox: The most immediate change they will notice is the introduction of installation codes. During installation, an installation code must be entered. These codes are generated by Red Hat. Depending on what services you purchased, they unlock different RHEL 5 components such as virtualization and high availability. RHEL 5 also uses new RPM GPG keys. If upgrading, you will need to install the new keys before updating packages.

The next big noticeable change is the move to YUM for software management and maintenance. In previous versions of RHEL, the up2date command could be used to download and install software updates from Red Hat Network. It has been replaced with the YUM command, and all the up2date graphical program has been replaced with a graphical program that use YUM as well. If you schedule package updates and installation with the RHN Web site, you will not notice a difference.

Administrators must also consider whether to use the virtualization feature in RHEL5. (Editor's Note: Due to trademark issues, Red Hat currently refers to Xen-based hypervisor technology as virtualization.) If your hardware is supported and you have systems with enough resources to share between the host system and the virtual machines, virtualization can help reduce the number of physical machines that need to be maintained. If you are a hosting service or allow your customers shell access to their Web site, virtualization can be beneficial.

Administrators should read the release notes for a complete list of caveats and known issues with RHEL 5. For example, dual-booting with Windows Vista requires additional steps for GRUB to boot Windows Vista.

Has SELinux become any easier to use with RHEL 5?

Fox: As with any developing open source technology, SELinux is improving with each release and version of Red Hat Enterprise Linux and Fedora. Of course there will be IT managers who continue not to use SELinux, but the version in RHEL 5 has improved significantly.

With the introduction of the SELinux Management Tool and the SELinux Troubleshooter, it is much easier to determine which services and programs are protected by SELinux. It is also much easier to customize the policy without having to write your own policy. SELinux booleans are used for this and can be enabled and disabled with the SELinux Management Tool.

The SELinux Troubleshooter alerts administrators of when a service has failed to start due to SELinux. It also provides suggestions on how to work around the issue without just disabling SELinux protection for the service.

How has the typical Red Hat Linux administrator's job changed with the introduction of Xen to RHEL 5?

Fox: With the introduction of virtualization in RHEL 5, administrators must now consider whether to continue using individual physical servers for each operating system instance or whether the performance of a virtual machine (VM) will suffice. Certainly, Web hosting companies who offer dedicated Web servers will benefit from virtualization since each client can have their own IP address and guest OS. The VMs on a host system can't access the data on the other VMs or the host, so client data is separated, yet one physical server can serve multiple clients.

IT managers should know that there are hardware requirements for virtualization. Currently, virtualization on RHEL 5 is only supported on x86 and x86_64 systems. It is offered for the Itanium2 but only as a technology preview. In addition to the obvious need for enough processors, memory and disk space to dedicate to each VM, the processor must have Physical Address Extension support. To use full virtualization -- the guest OS does not have to be aware of the VM layer -- the processor must be 64-bit and must have a hardware virtual machine layer.

An eWEEK review of RHEL5 said the Xen support is "half-baked" -- what are your thoughts on this?

Fox: As with SELinux, you will see the virtualization support improve over time. The basic functionality is there. You can create a virtual machine, install a guest OS on the VM, and connect to the guest OS as if it is a OS running on a separate set of hardware. The VM has its own IP address, doesn't have access to the data on its host system, and so on.

There are still parts that can be greatly improved, as with any software when it is first introduced. Because it is open source, it is important to provide these types of new technologies so that they can be used by the masses and so other developers and companies will get behind it. A few of the known issues such as the virtual Machine Manager not being able to start the VM are listed in the RHEL 5 release notes. However, this particular issue has a workaround by starting the VM from the command line first. This issue is also discussed in my book.

Have a question or a comment about the article? Email Jack Loftus, News Writer or visit our new blog at the Enterprise Linux Log.


Pre-loaded Linux = Increased adoption?


> QUESTION POSED ON: 25 April 2007
Do you think Dell's idea to pre-load Linux on PCs will increase adoption or will people stick with what's familiar (e.g. Windows)? How can they overcome the usual hurdles of driver incompatibility and hardware issues?

> EXPERT RESPONSE

I think Dell pre-loading Linux will have a significant impact on Linux adoption, although not necessarily because so many end users comfortable with Windows will decide to switch. Rather, Dell's involvement will motivate many hardware manufacturers to make reliable Linux drivers available, the lack of which has significantly retarded the adoption of Linux by many individuals and organizations interested in trying out Linux.

By reducing the barrier of drivers and thereby enabling these trials, Linux will start to achieve some momentum, which will in turn increase its adoption rate. So, I applaud Dell for moving forward on this initiative.



Microsoft Windows ousted at California school district


By Jack Loftus, News Writer
28 Feb 2007 | SearchOpenSource.com

By all appearances, the migration from Microsoft Windows to Novell SUSE Linux on the server and the desktop at the Windsor Unified School District in Northern California has been almost as pain-free as any IT professional could hope for.

By this summer, all 5,000 students and 250 teachers will be working off of a Linux-based thin client running OpenOffice.org, and the majority of the district's servers will be running Novell SUSE Linux Enterprise Server.

Closing Windows

Heather Carver, brought on as the director of technology and information services at Windsor in August 2006, said her hiring was a result of the district's desire to vet alternatives to a Windows upgrade for its network of seven schools.

When Carver arrived, Windsor had an aging Microsoft Windows environment running on 70 Hewlett-Packard and Dell servers spread across seven schools. Upgrades would have meant purchasing more powerful hardware and additional licensing costs totaling $100,000 – far too expensive for their limited IT budget, she said.

That said, Carver still investigated ways to remain on Windows, but an upgrade proved impossible. "We looked at keeping the physical environment, and how we could accomplish that. But in that scenario, if we could afford the software upgrades, then we could not afford the new hardware required to run it and vice versa," she said.

With Windows ruled out, Carver decided to standardize on SUSE Linux Enterprise Server with thin-client desktops running SUSE Linux Enterprise Desktop. Novell won out because of what Carver perceived as better support for the thin client desktops she planned to use in the classrooms and faculty areas.

Using Wyse terminals, the district also planned to offer students and staff thin-client desktops running OpenOffice.org. Ericom software – a Citrix alternative -- enabled the terminals to run the district's existing and irreplaceable Microsoft Windows educational applications, including Type to Learn, Reading Counts and Kid Pix.

The migration officially began in January at the Brooks school building, one of seven schools in the district that would serve as the testing environment for the project. ZENworks, Novell's systems management product, was also installed to serve as a central management and monitoring hub for the other six sites when they go online this summer.

Identity hiccup

Windsor did go over a few speed bumps in the migration to Linux involving managing user identities and authentication. For Carver, the ID issues were not debilitating enough to deter her from making the switch from Windows to SUSE, but the issue was still noticeable.

Specifically, custom scripts had to be developed to perform tasks such as granting faculty access privileges to student grade information, for example. These tasks had to be completed locally on a server-by-server basis. "We had to add all the system's users one-by-one," Carver said. "That's a pain when you have 500 users."

This problem was especially apparent on the thin clients she had set up in classrooms and teacher workstations. "At our Brooks facility, the OS wasn't that easy to manipulate for our thin clients," Carver said. "I love the OS, but it is not built for the thin client."

Carver said Novell promised more identity support during the summer in the form of a software update.

Linux thrives

Ultimately, though, the migration from Windows to Linux progressed quite smoothly.

"Each site has a server room with 10 servers," Carver said, for a total of 70 servers running applications like file and print, plus Novell's ZENworks and GroupWise for collaboration.

In particular, implementing ZENworks as a help desk for teachers and staff has resulted in a 90% reduction in the amount of time it takes IT staff to resolve problems. "We basically had no help desk before, but in the new system, each user has a distinct name and log-in credentials. Most places have that already, I know," she said.

The new setup also allows for better remote management. "[With Windows] we had spent half our time driving around; we had to touch every machine," Carver said. In a school system like Windsor, all that driving was costing an already strapped IT department too many resources.

Carver said it cost the district about $2,500 per school to migrate to Linux, compared with the estimated $100,000 it would have cost to upgrade their Windows infrastructure. In addition, buying more Microsoft Office licenses would have cost the district $100 per license, she said, whereas OpenOffice was free.

Linux as a learning tool

Ultimately, moving to Linux has enabled the Windsor School District to build out technology capabilities that wouldn't have been possible with Windows.

"[The students] are able to do more because Linux cost less," Carver said. "Our new computer lab [at Brooks] was set to cost $35,000 and ended up costing us $16,000 with Linux [on thin clients]."

And the kids love it too. "The kids think Linux is cool because it's new, but what they're really doing is stepping into the 21st century," Carver said.