To those who are interested in the topic of Linux Storage Management, planned for the 3/2009 issue and hitting the shelves July-September is my article of the same name in Linux+ magazine. I do not know how 3/2009 equates to July-September, but that is what I have been told. It is a 7 page article and gets into some great detail with storage management.
Recently I have been reading articles about the new Btrfs and its benchmarks for the Linux kernel. I have also been reading other articles relating to file system maintenance. At this point, regarding the article on file system maintenance, I am going to have to say “who cares.” According to the VAR Guy, Diskeeper is considering supporting Mac OS X and Linux. At this point in computing this support becomes increasingly insignificant and I will tell you why. Diskeepers’ target market caters to the hosting and virtualization end of computing. Their idea is to offer disk defragmentation software. But why? I can see how this may be optimal in a Microsoft Windows environment as the NTFS file system is a horribly designed file system.
Let us say that very soon, people are going to start trusting Btrfs. We know that Btrfs’ primary focus is server side hosting. It comes equipped with its own defragmentation tools that run both online and offline. So why would I need support for anything else? XFS has been out for many years and has gained the trust of storage administrators worldwide as a stable, excellent performing and extremely scalable file system solution. XFS offers both online and offline defragmentation. AdvFS supports online defragmentation and recently I have even heard proposals for an Ext4-fs online defragmentation implementation.
My point is: an intelligent storage administrator will not utilize a file system that is not appropriate for his/her environment. This may end up being a losing battle for Diskeepers; if they venture outside of the Microsoft world and especially into the world of GNU/Linux.
Times are changing and what is classified as mid-size to enterprise class computing is not what it used to be. File storage and virtualization concepts have evolved where it becomes increasingly difficult for 3rd parties to step in and provide additional tool sets to reap from the glory that the developers and maintainers of such solutions deserve. The communities developing the latest and greatest solutions are finally realizing that to provide a complete solution is the only way to go.
Also who knows what will become of other 3rd party utilities such as file shredders over snapshot-enabled file systems and volumes. My advice to Diskeeper is: “Stick with what you know.”
As I had mentioned in my last post, I was planning to upgrade from Sun’s VirtualBox 2.1.4 to 2.2.0. You can read my full review of 2.1.4 here. Well, I installed it over the weekend but I have not had much time to play around with it. Although I will say that the upgrade went smooth. Note that this was via rpm installation through Fedora GNU/Linux. The upgrade displayed no obvious problems, so far; not that I really expect any. All my virtual client profiles had to be converted over to a “new” format so that VirtualBox 2.2.0 could play nice with them.
My main objective to testing out 2.2.0 was to find out if there are any improvements in performance; especially when virtualizing OpenSolaris 2008.11. It was kind of hard to tell. I want to say that I noticed a slight improvement but it was not significant enough for me to get excited about. Not too long ago and very briefly I had installed VirtualBox 2.1.4 for Windows XP on a computer at work, it too didn’t perform as well as I had hoped. When I was working with some LiveCDs through it (i.e. Fedora 11 and PuppyLinux) it virtualized the operating systems a little slower than I would have expected. I should see if there has also been improvement there (unless someone beat me to it, then please share your opinions). As I personally do not use Microsoft Windows anymore, this is the least of my concerns.
Other than the performance of Sun’s OpenSolaris operating system, I am still not disappointed in VirtualBox and continue to use it.
Some of you may remember my review on VirtualBox 2.1.4. Sun Microsystems just released VirtualBox 2.2.0; available for download here. According to their Changelog for this release, some major bug fixes and newly implemented features are integrated into this major release. Although I do not see it specifically called out, I do hope that they addressed certain performance problems (excluding performance fix to hypervisor in context switching), specifically when virtualizing Sun’s OpenSolaris 2008.11 from my Fedora operating environment.
I will be sure to update from version 2.1.4 to 2.2.0 and write a quick review if anything is worth noting on top of the last one written.
For the past month or so I have been reading of the potential acquisition of Sun Microsystems by IBM. I have seen this dubbed as the Blue Sun. After a month of negotations it now looks like this may be a reality. What does the future hold if and when this pushes through? What are we to expect?
For years, I have been using Sun and IBM products. And coming from the data storage industry, I have also had the great opportunity of seeing where these systems get deployed. Sun’s Solaris UNIX operating system and related intellectual property (IP) have long been popular in the states. While their influence is world wide, I have noticed more IBM AIX deployments in Europe. Note that these comments reflect my personal experience as opposed to the reality of the situation.
While they have made great milestones in the development of AIX, IBM’s operating platform still seems somewhat arcane compared to the Sun Solaris counterpart. And while IBM offers their own UNIX OS, they still do an excellent job of supporting GNU/Linux.
I do not know what IBM’s intentions are if/when the acquisition is complete, but with the newly acquired IP, we may be seeing a lot of changes to both UNIX and Linux platforms. Will we see ZFS or DTrace relicensed under the GPL? Don’t know, it is still too soon to tell. This may also just be wishful thinking on my part.
I can say this: Microsoft may face a little more of a struggle in the enterprise computing arena. This may be especially true since the older standing enterprise class UNIX providers are slowly disappearing from the map. As is the case with SGI being acquired by Rackable Systems.