Whither Sun Microsystems?

The recent fourth quarter reports from server manufacturers was dim, and Sun Microsystems was by far the worst (with a 35% loss compared to the same period last year). On top of this, Sun just announced in October (within their 8K filing for the SEC) intentions to lay off 3000 employees in the next 12 months. Infoworld also had a nice piece on this; according to Larry Ellison, CEO of Oracle, the company is losing $100 million each month the European Union regulators put off accepting the merger.

With the Oracle acquisition in progress, there are a lot of questions about the future viability of Sun Microsystems, and of some of its products.

I don’t think people realize just how important the Sun group of products are, and what an impact it would have if most – or even some – of the products were cancelled. Consider this list of Sun products:

Most of the most popular products were mentioned by Oracle in their Sun Acquisition FAQ (PDF), stating that they will increase money spent on each over what Sun spent. These products include: Java, Solaris, SPARC, StarOffice, NetBeans, virtualization products, Glassfish, and MySQL. Other products were not mentioned – such as Lustre, the Modular Data Center, and others.

The list above also does not list the technologies that were spearheaded by Sun – and some still are: ZFS, NFS, NIS (and NIS+), dtrace, containers, and smc.

It would be unfortunate – and materially significant – if Sun were to go under or if any of the majority of their products were to be cancelled. One can only hope this does not happen…

Converting OpenBSD 4.1 guest from VMware to VirtualBox

This turned out to be easier than it would appear, although the vmware-tools needs to be extracted from the system.

The first thing that I did was to add the virtual hard drive created by VMware to the list of hard drives that VirtualBox makes available. This is in the Virtual Disk Manager from the main menu (Ctrl-D).

VirtualBox has native support for VMDK disks, the format that VMware uses. However, the documentation suggests there are restrictions, although the documentation may be obsolete: for one, it appears that snapshots are now possible, although the documentation suggests otherwise.

Having added the disk using the Virtual Disk Manager, I then created a new virtual environment and used the disk instead of creating a new disk from scratch. The disk was picked up and used seamlessly.

However, booting the environment (predictably) had problems: the VMware root disk was /dev/sd0a, but the VirtualBox root disk was /dev/wd0a. Thus, everything was fine until /etc/fstab was read, then OpenBSD presented the option to utilize a shell to fix the problem.

At the shell, it was necessary to mount the root filesystem read-write:

# mount -o rw /dev/wd0a /

Then editing /etc/fstab to use the correct disk was all it took.

However, VMware does not use an OpenBSD package to install the software, and apparently just drops it into the environment – and not in /usr/local either. All of the BSDs fiercely recommend placing every addition to the system in /usr/local – every add-on package does, from BIND to PHP to Apache to KDE to OpenOffice – everything. So for VMware to litter across the filesystem in this manner is very bad taste – and even without a package to extract it from the filesystem properly.

However, using locate, we can find the vmware-tools (or what looks like all of them):

# locate vmw

These files and directories can then be removed, although if the disk is to be used by VMware again you may not want to. However, using a virtual environment in two different products on a regular basis sounds like a recipe for disaster.

Running Kubuntu Intrepid Ibex Alpha 6 on a Compaq nc4010

e1000e driver in most recent Linux kernel causes corruption!

Before you try using Intrepid Ibex Alpha 6: there seems to be a problem with the e1000e driver that causes the hardware to be corrupted and could render the e1000e card useless – and even unrepairable. Even if you are using a Linux system that currently uses an e1000 driver, the new Linux kernel shifted some of the e1000 support to the e1000e driver. If you are using the Compaq nc4010 as this article describes, you should be fine: the nc4010 uses the tigon3 driver. There is a bug entry in the Ubuntu bug lists, and the linux-net mailing list has a thread on the bug. It also would appear that a fix went into the -mm kernel tree (a recognized Linus tree spinoff) as of 2.6.27-rc5-mm1. If you are hardy enough to run Ubuntu Intrepid, perhaps you could tangle with the experimental -mm kernel as well – and sleep better at night knowing your hardware won’t be wasted.

This turned out to be quite a challenge. Firstly, there was no way to install it directly – the previously mentioned Billix didn’t accomodate Intrepid and I didn’t have any large enough USB sticks to put a USB bootable image onto – assuming there is one for Intrepid.

However, remaining undaunted, I was able to install Intrepid without too much trouble – though any nontechnical user would have been stopped right up front. How’d I do it?

First, I installed VirtualBox onto another available system – using VirtualBox 2.0 – and then downloaded the Kubuntu install CD to that system. Installing Kubuntu Intrepid Ibis to VirtualBox was not a problem; everything went well. It was, however, quite slow! It turned out I had much less memory in the system than I thought – so between the 512 Mb in the system and the 300+ Mb that was allocated to the VirtualBox instance, I did a lot of waiting (sigh).

I made sure that the hard disk that was created in the virtual environment was smaller than the actual disk used by the system I was going to put Intrepid onto.

Once the virtual environment was complete and the install was finished, I stopped the environment and reconfigured for less memory. I then restarted the environment with a DSL disk (so as to not use the created virtual disk in any form).

I extracted the hard disk from the Compaq nc4010 that was to have Intrepid installed onto it, and removed it from the hard disk cage that notebooks like to use. I then connected the hard disk to a USB port using a cable adapter.

Now – with the virtual environment running DSL and a unused disk configured with Kubuntu Intrepid, and the host system running OpenSUSE 10.3 with the target disk attached via USB, the only thing left was a disk copy over the network. Using nc (or netcat) permitted the copy going direct from the virtual guest to the virtual host.

The networking had to be set up, and required the bridged mode. It appeared that the host had to already be configured for networking and active in order for the guest to be able to talk to the host (using NAT), but perhaps that was just me.

Once networking was set, the only thing that was needed was to copy from the guest:

# nc -p 4117 < /dev/hda

And copy to the host:

# netcat -l -p 4117 > /dev/sda

Note that this copy will copy the entire disk, including partition tables. This was by design and worked fine (apparently).

The biggest problem (aside from speed and memory overcommitment) was the fact that nc did not stop after copying all of the data from the hard drive – and there was no visible progress report anywhere. The destination disk had no activity light; nc had no progress report available; and so forth.

Another “problem” was the fact that the command nc did not exist on OpenSUSE but did in DSL: OpenSUSE used the full name for the command, netcat.

Once everything was copied, I could shutdown the virtual guest environment, and put the hard drive back into the original host (using the cage as mentioned before).

This did work and worked beautifully. The biggest problems came not from the installation, but from the fact that the installation is Alpha 6. Some of the programs on the task bar don’t have a proper background, several things crashed, icons are missing for some programs in the menu (including some programs that have icons shown elsewhere, like Amarok). I don’t know whether to be aghast or just patient – it is alpha software, after all. I’d just expected the most obvious bugs to be gone, but whatever.

Lastly, every time I hear that name…. am I the only one who thinks of the man called Intrepid?

OpenSolaris 2008.05

OpenSolaris 2008.05 (for x86) was released earlier this month, along with a new OpenSolaris web site (which to me bears more than a passing resemblence to the OpenSUSE web site). I don’t know the difference between that web site and OpenSolaris.org, but somebody must think its worth it.

I’ve already ordered my CD of OpenSolaris 2008.05; now to find something to run it on. Alright already, so that’s backwards – but I’m always trying the new UNIX versions. And why not? This is an excellent way to find out what is likely to be in Solaris 11.

The CD of OpenSolaris 2008.05 is also a live CD, which means you can try it out anywhere the operating system can run: just reboot with the CDROM in the drive.

Jason Perlow of ZDNet wrote a nice review article titled OpenSolaris: What Ubuntu wants to be when it grows up on May 5, 2008. It isn’t a step by step technical review, but it is good nonetheless. Another review was recently written by Ashwin Bhat K S about his experiences with OpenSolaris 2008.05. He used a laptop, so this review is doubly interesting. There is another review by Milind Arun Choudhary which talks about installing OpenSolaris (Nexenta, actually) into a VirtualBox virtual machine under MacOS X.

I can only hope that OpenSolaris driver support will improve; last time I installed it onto a current laptop it had no wifi connectivity and no network connectivity because of missing drivers. Is it any surprise then that Ashwin’s first recommendation is more network drivers and more wifi drivers?

Update: Another person has done up a nice article about installing OpenSolaris 2008.05 – their first and only post apparently!