I’ve made a few additions to the original post about hacking your Hardy Heron to accept Java as it stands. Check it out and read the comments too. Jazz has added a lot to what I’ve said. Don’t forget to visit the forum thread that started this all.
NOTE: this tip may be specific to Ubuntu, as far as the command goes…
One thing that I’ve noticed is that many distros are starting to change their partition references in /etc/fstab to a UUID from the more familiar (but no friendlier) reference to the volume in /dev (such as /dev/hdc3, for example). This might be a sign of things to come, since any distro based on the 2.6 Linux kernel has pretty much dropped the /dev for udev (hot plugging for cold plugging, etc). After looking into things, its good for situations where drives are being moved or removed now and then (as in when you plug in your USB hard drive).
However, if you’re a constant alpha/beta tester at heart, like I am, you are probably using a main Linux partition or drive to work from and have either a spare drive or partition that you frequently blow away and replace what’s on it with the latest fun toy, as I do. I don’t know how many distros are doing this with the UUID so far, but I suspect that Fedora does and I know that Ubuntu and Mandriva are now doing it. So, what happens if you are tired of what you’ve been testing and you want to install something new? As soon as you format the drive, the UUID will change. How do you reference the drive in your /etc/fstab, so that you can either find it when you want to mount it, or when you want to boot from the new distro? The first “U” is for “unique”, and they are! Here’s what’s currently in my /boot/grub/menu.lst for the Mandrake alpha 2 that I’m playing with: root=UUID=b414e306-0582-4572-926d-a8c113bf34bb.
That’s not exactly the stuff of memorization, and, in a few hours, it’ll be different, as I’m going to hose it and reinstall it with a totally different configuration (I wanna try KDE 4!!!). So, how do I get the new volume ID to put in /etc/fstab and /boot/grub/menu.lst? With Ubuntu, you have a command called vol_id that you can run as root to determine various things about the volumes that are accessible to your computer.
The command is vol_id. So, the command to find a volume ID is:
sudo vol_id -u device
where device is the /dev entry for the partition you want to know about. for example,
sudo vol_id -u /dev/hdc3
would yield the long scary output that you need to enter into your files.
I hope that this helps all you distro hoppers out there as much as it has me.
Have you done the nutso and “upgraded” or tried out the development release of Ubuntu’s Hardy Heron? I did. I decided to use the command “update-manager -d” to run Gutsy Gibbon’s Update Manager with the flag to update the distribution to the next release. Currently Hardy is at 8.04 alpha 3. For those Debianites, this would be like changing your /etc/apt/sources.list file to point to the “testing” sources and doing the “apt-get dist-upgrade” command after “apt-get update”.
The upgrade went reasonably well with only one temporary show stopper. Java GUI apps stopped working when I upgraded. It seems that there’s a bit of a bug in Sun’s Java 6 that doesn’t agree with the new x.org server (from what I gather). They will fix it in Java 7, which will be a while yet. Thankfully, I found this ubuntu forum post, that helped me solve the problem. I used the sed line:
$ sudo sed -i ’s/XINERAMA/FAKEEXTN/g’ /usr/lib/jvm/java-6-sun-1.6.0.03/jre/lib/i386/xawt/libmawt.so
I hope that this will help you too!
ADDENDUM:
- One thing that you will find is that after every Java update and many X.org updates, your java will stop working again. Adding the “replacement” string “FAKEEXTN” again should fix the behavior.
- As noted by jazz in the comments (Please read the comments on this post), there are different paths for x86 and x64.
- You may or may not require the additional hacks to the /usr/bin/<java application> script to make it work. I didn’t for Azureus or Frostwire.
- Any time someone says to “paste” or write a command with the SUDO command or to do “as root“, think before you copy
- Read the MAN PAGE for SED (and AWK) before using it. Its a powerful command.
Addendum 2: (this is a short one)
Don’t expect anything big with Firefox for this one, at the moment, as Firefox 3 is the designated upgrade from the 2 branch (whether I approve or not - I’m testing and submitting bug reports) ad there seems to be a problem with java and FF3. If you know how to get it working, let me know. There are a few little things about FF3 that are great, but, mostly, I hate FF3. It sucks the biggest moose appendage at the moment. I yearn for the simple days of Phoenix 0.1…
Run, Skye!
I’ve not quite done this before, so I wasn’t sure if the rules were the same or not. There have been times that I’ve relied on chroot to fix a dual-booted machine, but not to install an OS from a booted Operating System.
So, what I did was install Gentoo on a spare hard drive from my existing OS, Ubuntu. I used the same method that I would have had I used the minimal installer, which seems to be the only method that works, despite the graphical installer on the 2007.0 CD.
What prompted me to do this? I’ve missed the snappiness that I had with my Gentoo system before (this system, actually) and i wanted to try Gnome on Gentoo, as I’ve never done so. I did try the CD installer but it really messed things up big time, and the screen was nearly unusable on that super slow frame buffer driver that the Live CD used. So, surfing the net, etc, was unbearable while doing the install. Once installed, X was not working (apparently the video drivers aren’t even on the CD) and the system was full of unneeded packages. The only way that I’ve been able to quickly get a Gentoo system up and running was to do it from the command prompt, which was fine with me.
After finding a little tidbit in the Gentoo Alternate Install Doc regarding installing from another operating system. It turns out that if you have the normal toolchain required to compile code and the chroot tool, you’re in business.
This is a basic install so far, just to see if I remembered the right modules to compile into my kernel. I could have used Genkernel, but I wanted to go leaner than that. My old P-III doesn’t have SATA or SCSI, for example. So, I didn’t compile them in. I know that there’s still more to yank out, but I wanted to get it up and running first.
I chose to use the bootloader that I had installed already (which was actually on the drive I installed Gentoo to), so I added a stanza to my /boot/grub/menu.lst to include Gentoo and I was off.
Perhaps, I’ll update the news when I have more installed. I’m curious as to how much I can do in the chroot.
It looks like the Windows machine has died for good. I was trying to swap over files from my Debian install to the Windows machine so that I could wipe my drive and replace Debian with Ubuntu. About 1/2 way through the file transfer (over 100mbps ethernet, which is never any where near that, ever), the Windows machine unexpectedly rebooted. It stopped at the spot where Windows would begin to load, complaining about an unrecoverable error and that I was to hit CTRL+ALT+DEL. This did not help.
Under the Recovery Console, which I go to by insterting my Windows CD and booting from it, I did a CHKDSK and it also complained about an unrecoverable error. A DIR request yeilded no directory but a complaint about an unrecoverable error.
Without any other recourse, I am going to try and reinstall Windows. So far, after it prechecked the “not quick” install, the installation has been stuck at 39% for the last 45 minutes. So, its trying to do a bad block check and work around those, I guess. Its not doing it very quickly, though.
I guess that I’ll have to admit it and work out getting another drive installed. The DiamondMax 8 that is in there is the same drive that’s installed in the workstations at work, and our IT guy has replaced most of them already. Its just time, I guess. Time to play Taps.
Sometimes I wonder why I bother playing around, as I just go back to the old standard, for the most part. Okay, that’s not really true. I’ve installed the network install version of Debian 4.0R1 (Etch). I also don’t know why I did it this way, as I just downloaded a few metapackages and ended up with a reasonably full system anyways. Interestingly, there is no GUI instant messenger client installed, though. When I searched for it (looking first for Pidgen, as the “Desktop” metapackage installed Gnome), I found Gaim, which has since been renamed to get AOL off their backs.
I guess that this isn’t that important, as I’ll probably hose this install anyways when my 80G drive arrives. I thought that I’d see what was up with the “base” for alll those popular distros out there, since I hadn’t kept up with it for quite some time. I do notice one thing so far that I really don’t like and that’s the fonts in Epiphany, the Gnome browser. They’re terrible. Maybe it’ll be back to Ubuntu again after all.
