I’ve deployed a number of VMware installations for clients and have been reasonably happy with the results. I recently spoke with someone who advocated using Xen for GNU/Linux servers and Windows Virtual Server for Windows servers. I would agree that in an exclusively GNU/Linux environment that Xen is my preferred solution. However, most of my clients need both Windows and GNU/Linux servers and I would prefer them to be hosted on the same server. I also think it’s advantageous to use only 1 virtualization solution for an organization, especially smaller organizations. At home I’ve been using VirtualBox OSE (even though I’d qualify for the full version under the Personal Use license). I’ve been considering why it might be a better choice than VMware for businesses.

I’ve been meaning to write about this for a while and since it’s come up recently (Write Intent Bitmaps, Software RAID) I figured now would be a good time. Basically what the write-intent bitmap does is keep track of which blocks are being used so that in the event that a resync is needed (power failure, kernel crash, disk temporarily offline) only the blocks that are out of sync are copied. This drastically improves resync times. Of course there is a downside.

I had some trouble restarting a xen user domain tonight. The shutdown, destroy, console, and list xm commands were all simply hanging. I finally tried restarting the xend service which fixed the hanging problems. But when I tried to start the user domain that originally gave me problems I got this error message:

Error: Device 769 (vbd) could not be connected.
Device /dev/vg/lv is mounted in a guest domain,
and so cannot be mounted now.

Over the past few weeks I have increasingly found myself unhappy developing Rails applications in NetBeans 6. While I was initially happy with the productivity boost from moving from writing applications in nano to using a full IDE I found myself reverting to using nano. The reason was simply that NetBeans was putting too much load on my desktop system. The system isn’t exactly slow (Athlon 64 X2 4600+, 2GB RAM) yet when running NetBeans the system load was hovering between 4 and 5 making development tedious. Before trying NetBeans I had read a lot about various IDE’s and I decided to revisit that research. One consistent complement to Eclipse with RadRails was that the performance was snappy. I decided to give it a try, if for no other reason than to expand my Rails IDE experiences.

I've been noticing that between desktops, laptops, and servers that there have been a lot of updates over the last week. I've seen updates for KDE packages, CUPS, Avahi, PHP, PostgreSQL, and more. Some are security related and others I believe come from the gutsy-updates repository and are just newer versions. In any case it seems like a good time to check in on any systems and update them.

I was working on my laptop and realize that I was seeing the errors “device-mapper: ioctl: error adding target to table” and “device-mapper: table: 253:1: linear: dm-linear: Device lookup failed” flood my logs. My CPU usage was also extremely high. Apparently this is an issue with EVMS and recent Linux kernels. There are a couple ways to resolve this.

Traffic shaping is one of those topics that most people find mystifying when they attempt to research it. I’ve been doing traffic shaping work for years and I still find myself learning new things. I thought with some recent work I’d write down a few notes on practical setups for Linux using the tc (traffic control) command. I also want to highlight some common issues people need to address when designing traffic control systems.

For better or worse I’ve been needing to check drives for errors quite a bit recently and have been using the badblocks command to do it. It’s definitely a command that people should become familiar with. Where fsck checks the file system for problems, badblocks attempts to ascertain the status of physical media which can include hard disks, usb drives, and other usb devices such as media players. I’ve compiled a detailed list of commands I’ve been using.

There's something you don't see a lot of: an openssh vulnerability. Time to patch those clients. Details can be found at: USN-566-1: OpenSSH vulnerability. The vulnerability has to do with X11 forwarding which I always disable so I assume my systems are safe but I will still upgrade promptly.

Single sign-on is the desire to have 1 set of authentication credentials and account information per user on a network. It’s a problem with which institutions large and small struggle. For large institutions the problems are often related to scalability and compatibility with large numbers of disparate systems. For smaller organizations the issues are often related to the manageability of the infrastructures. Finding a solution with the right fit for a particular network can be daunting to say the least. I decided to review what the options were to consider implementing one of them for a network with about half a dozen users and about 50 systems (many virtualized).