The Weezey Geek

How To Migrate VMWare Server virtual machine to VirtualBox Headless

2012-02-28T12:09:00.001-05:00

Visualization has changed the way we use computers and servers. Now when I'm shopping for new hardware I always make the conscious decision to select an extra couple of cores or 4GB of RAM so that the computer I'm building can support running some extra virtual machines. I spent most of my time using VMWare's Virtual Server which became VMWare's free product and now I've moved to using VirtualBox which seems to leave me with much more responsive virtual machines and can also be run in a convenient headless mode on my Linux servers.

First let's install VirtualBox...

# sh VirtualBox-4.1.8-75467-Linux_amd64.run
Verifying archive integrity... All good.
Uncompressing VirtualBox for Linux installation...........
VirtualBox Version 4.1.8 r75467 (2011-12-19T13:15:47Z) installer
Installing VirtualBox to /opt/VirtualBox
dPython found: python, installing bindings...
Building the VirtualBox kernel modules

VirtualBox has been installed successfully.

You will find useful information about using VirtualBox in the user manual
/opt/VirtualBox/UserManual.pdf
and in the user FAQ
http://www.virtualbox.org/wiki/User_FAQ

We hope that you enjoy using VirtualBox.

And for headless operation we need to install the extension pack

# VBoxManage extpack install /usr/src/Oracle_VM_VirtualBox_Extension_Pack-4.1.8-75467.vbox-extpack
0%...10%...20%...30%...40%...50%...60%...70%...80%...90%...100%

Now, since I'm booting a Windows XP guest, I need the drivers from Intel for the Intel PRO 1000 MT (82540EM) card, so download those and put them in a folder called IntelPRO then make it into an ISO so we can hand it to the virtual machine.

# mkisofs -V IntelPRO -o IntelPRO.iso -r IntelPRO/
INFO: UTF-8 character encoding detected by locale settings.
Assuming UTF-8 encoded filenames on source filesystem,
use -input-charset to override.
33.72% done, estimate finish Tue Feb 28 09:35:15 2012
67.48% done, estimate finish Tue Feb 28 09:35:15 2012
Total translation table size: 0
Total rockridge attributes bytes: 253
Total directory bytes: 0
Path table size(bytes): 10
Max brk space used 0
14832 extents written (28 MB)

Now let's build our new virtual machine to host our existing VMWare Server disk...

# VBoxManage createvm --name "New Machine" --ostype WindowsXP --register
# VBoxManage modifyvm "New Machine" --memory 1024 --acpi on --ioapic on
# VBoxManage modifyvm "New Machine" --nic1 bridged --nictype1 82540EM --bridgeadapter1 eth0
# VBoxManage modifyvm "New Machine" --vrde on --vrdeport 3330
# VBoxManage storagectl "New Machine" --name "IDE Controller" --add ide --controller PIIX4
# VBoxManage storageattach "New Machine" --storagectl "IDE Controller" --port 0 --device 0 --type hdd --medium /var/lib/vmware/Virtual\ Machines/oldvmwaremachine/oldvmwaremachine.vmdk
# VBoxManage storageattach "New Machine" --storagectl "IDE Controller" --port 0 --device 1 --type dvddrive --medium /opt/VirtualBox/additions/VBoxGuestAdditions.iso
# VBoxManage storageattach "New Machine" --storagectl "IDE Controller" --port 1 --device 0 --type dvddrive --medium /full/path/to/IntelPRO.iso
# VBoxManage startvm "New Machine" --type headless

Now we can open an RDP session to the server's IP on port 3330 and see the console. It's worth noting here that unless you have the ioapic enabled on VMWare disks, the system won't function correctly, also that this isn't very secure so you should have a firewall in place or use one of the documented authentication methods.

Once connected to the machine, you can use the VirtualBox Guest Additions image to install that and then use the IntelPRO image to install the drivers for the network driver. That's all there is to it and your VMWare machine is now a VirtualBox machine.

The absolute best guide for creating a git repository

2012-02-18T23:54:00.000-05:00

This guide is easily the best source for creating a git repository from a local folder. Unfortunately it was down today when I needed it so I've placed the important stuff below in case it happens again and I want it. (and don't want to use Google's cache)

I also want to be sure I omit any .svn or .DS_Store files...

# echo .DS_Store >> ~/.gitignore
# echo .svn >> ~/.gitignore

Change to the local directory which you want to create a repo in...

# cd myproject
# git init
# git config --global core.excludesfile ~/.gitignore
# git add *
# git commit -m "My initial commit message"

Now, create the repository on your Git server.

# cd /var/git/
# mkdir myproject.git
# cd myproject.git
# git --bare init

Now, go back to your local repository, and add the newly created remote repository so it tracks from the remote repository (origin).

# git remote add origin ssh://git@example.com:2227/var/git/myproject.git
# git push origin master

Now, to ensure that your local branch is tracking when you do a fetch, you need to use -f option to force a new local branch to be created even though it already exists.

# git checkout origin/master
# git branch -f master origin/master
# git checkout master

Kudos for the ignore stuff goes to this guide

Exchange 2010 SP1 Complains of WinRM Errors

2012-02-09T15:06:00.001-05:00

So after close to 24 hours of screaming at Exchange for being a dirty tit, I've got the solution to the problem on a customers' server.

The Problem
When opening Exchange Management Console (EMC) you get an error that reads:

"Connecting to remote server failed with the following error message: The WinRM client cannot process the request. It cannot determine the content type of the HTTP response from the destination computer. The content type is absent or invalid."

Most documents/forums/posts I could find pointed to these three fixes to that problem cited on technet

Remote PowerShell uses Kerberos to authenticate the user connecting. IIS implements this Kerberos authentication method via a native module. In IIS Manager, if you go to the PowerShell Virtual Directory and then look at the Modules, you should see Kerbauth listed as a Native Module, with the dll location pointing to C:\Program Files\Microsoft\Exchange Server\v14\Bin\kerbauth.dll. If the Kerbauth module shows up as a Managed module instead of Native, or if the Kerbauth module has been loaded on the Default Web Site level (instead of, or in addition to, the PowerShell virtual directory), you can experience this issue. To correct this, make sure that the Kerbauth module is not enabled on the Default Web Site, but is only enabled on the PowerShell virtual directory. The entry type of "Local" indicates that the Kerbauth module was enabled directly on this level, and not inherited from a parent.

If the WSMan module entry is missing from the global modules section of the C:\Windows\System32\Inetsrv\config\ApplicationHost.config file, as follows:
<globalmodules>
<add image="C:\Windows\system32\wsmsvc.dll" name="WSMan">
This will result in the WSMan module displaying as a Managed module on the PowerShell virtual directory.
To correct this, make sure that the WSMan module has been registered (but not enabled) at the Server level, and has been enabled on the PowerShell virtual directory.

If the user that is attempting to connect is not Remote PowerShell enabled. To check if a user is enabled for Remote PowerShell, you need to open the Exchange Management Shell with an account that has been enabled, and run the following query.

(Get-User <username>).RemotePowershellEnabled
This will return a True or False. If the output shows False, the user is not enabled for Remote PowerShell. To enable the user, run the following command.

Set-User <username> -RemotePowerShellEnabled $True

Though those are the common issues my IIS setup looked fine and none of those were out of order, but I found this guy who suggested the following:

After opening the web.config file and looking in the file carefully I noticed that the file has two environment variables used %ExchangeInstallDir% and %ExchangeInstallDirPath%. These variables are supposed to be registered in advanced system properties page of the operating system. In my case the physical path for %ExchangeInstallDirPath% was missing. This path should be the actual installation directory of your exchange server (Normally C:\Program Files\Microsoft\Exchange Server\V14\)

This also, didn't fix the issue for me, so I was trying to repair install Exchange (which you can't do), so I was going to uninstall and reinstall and when I attempted to uninstall (which you can't do either because the mailboxes are tied to the Active Directory) I got an error complaining about IIS being in the wrong mode.

If you have checked all of the above, the problem could simply be that IIS is running in 32-bit mode instead of 64-bit.

To change it, run the following: cscript c:\inetpub\adminscripts\adsutil.vbs SET /w3svc/AppPools/Enable32BitAppOnWin64 False

As soon as I did that the EMC and EMS worked correctly to connect to the local Exchange server.

OS X Lion 10.7.2 with NVIDIA GTX 570

2011-10-25T11:31:00.000-04:00

Disclaimer: This is essentially an update, so read this original post first if you haven't and once again, I'm not responsible if you cock up your machine, it comes with the territory. You'll need to learn how to de-cock your machine sooner or later.

I know, I know, I said I wouldn't update ever again, but I was having a weird issue with iTunes and the icons in 10.7.2's Launcher were bigger, I got dazzled. Anyway here are the fixes I needed to make to boot up my hackintosh with my EVGA NVIDIA GTX 570 with full resolution and openGL...

Using my system in the state from the initial guide, I installed the Mac OS X Lion 10.7.2 Update and rebooted. Results were expected, my NVIDIA GTX 465 worked perfectly, but my GTX 570 came up single monitor at 1024x768.

So, we need to make the same modifications we did previously to get things working with a couple little differences...

$ sudo su
Password:
# mkdir /Users/<username>/Desktop/Backup-GTX570-Mods-10.7.2/
# rsync -av /System/Library/Extensions/NVDAGF100Hal.kext/ /Users/<username>/Desktop/Backup-GTX570-Mods-10.7.2/NVDAGF100Hal.kext/

Now let's add our device IDs..

# nano -wT2 /System/Library/Extensions/NVDAGF100Hal.kext/Contents/Info.plist

Add your device ID to the IOPCIPrimaryMatch key list, mine is 1081 but yours may be 1086.

  <key>IOPCIPrimaryMatch</key>
  <string>
    0x06c010de&0xffe0ffff
    0x0dc010de&0xffc0ffff
    0x0e2010de&0xffe0ffff
    0x0ee010de&0xffe0ffff
    0x0f0010de&0xffc0ffff
    0x104010de&0xffc0ffff
    0x124010de&0xffc0ffff
    0x108110de&0xffc0ffff
  </string>

Next, install the July 22, 2011 version of ATY_Init.kext from netkas.

The updated version of GeForceBundle needs to be fixed. This time we're replacing the following hex codes:

eb a8 83 f8 02 7c 15
will be replaced with
eb a8 83 f8 03 7c 15

78 e8 83 f8 02 7c 11
will be replaced with
78 e8 83 f8 03 7c 11

Back things up first...

# rsync -av /System/Library/Extensions/GeForceGLDriver.bundle/ /Users/<username>/Desktop/Backup-GTX570-Mods-10.7.2/GeForceGLDriver.bundle/

You can do this fix using this helpful perl command line script.

# perl -pi -e '$c++ if s|\xeb\xa8\x83\xf8\x02\x7c\x15|\xeb\xa8\x83\xf8\x03\x7c\x15|; s|\x78\xe8\x83\xf8\x02\x7c\x11|\x78\xe8\x83\xf8\x03\x7c\x11|; END { unless ($c) { warn "No substitutions made\n"; $? =1 } }' /System/Library/Extensions/GeForceGLDriver.bundle/Contents/MacOS/GeForceGLDriver

A step I didn't do previously is to modify libclh.dylib to allow OpenCL 2.0. It shouldn't be necessary for GTX 570 cards but for complete hack-ness, I opted to do it anyway...

8b 87 1c 0c 00 00 89 06 8b 87 20 0c 00 00 89 02
will be replaced with
31 c0 ff c0 ff c0 89 06 31 c0 89 02 90 90 90 90

# perl -pi -e '$c++ if s|\x8b\x87\x1c\x0c\x00\x00\x89\x06\x8b\x87\x20\x0c\x00\x00\x89\x02|\x31\xc0\xff\xc0\xff\xc0\x89\x06\x31\xc0\x89\x02\x90\x90\x90\x90|; END { unless ($c) { warn "No substitutions made\n"; $? =1 } }' /System/Library/Extensions/GeForceGLDriver.bundle/Contents/MacOS/libclh.dylib

Now we edit the AppleGraphicsPowerManagement.kext to add power management for our card. Once again our system is set to be an iMac 12,2 so that's where I've added the following (in black)...

        <key>iMac12,2</key>
        <dict>
        ...
          <key>Vendor1002Device6740</key>
          <dict>
            <key>control-id</key>
            <integer>17</integer>
            <key>max-power-state</key>
            <integer>2</integer>
          <dict>
          <key>Vendor10deDevice1081</key>
          <dict>
          <key>Heuristic</key>
          <dict>
            <key>ID</key>
            <integer>0</integer>
            <key>IdleInterval</key>
            <integer>250</integer>
            <key>SensorOption</key>
            <integer>1</integer>
            <key>SensorSampleRate</key>
            <integer>4</integer>
            <key>TargetCount</key>
            <integer>5</integer>
            <key>Threshold_High</key>
            <array>
              <integer>93</integer>
              <integer>93</integer>
              <integer>93</integer>
              <integer>100</integer>
            </array>
            <key>Threshold_Low</key>
            <array>
              <integer>0</integer>
              <integer>97</integer>
              <integer>97</integer>
              <integer>100</integer>
            </array>
          </dict>
          <key>LogControl</key>
          <integer>1</integer>
          <key>control-id</key>
          <integer>18</integer>
        </dict>
        <key>default-control-id</key>
        <integer>18</integer>
      </dict>
      <key>iMac 9,1</key>
      <dict>
        <key>IGPU</key>

Now we rebuild the caches..

# rm -rf /System/Library/Caches/com.apple.kext.caches/Startup/kernelcache
# kextcache -system-prelinked-kernel
# kextcache -system-caches

And I needed to update my EFI string, I had to remove a section because it wasn't being recognized or something. Here's my updated org.chameleon.Boot.plist

Once I rebooted with that, all was perfect but then I figured out what I'd screwed up and so I was able to combine my EFI strings using OSX86Tools so I updated my EFI string to have both video cards in it and an ethernet card...

<plist version="1.0">
<dict>
  <key>Kernel</key>
  <string>mach_kernel</string>
  <key>Kernel Flags</key>
  <string>-v arch=i386 npci=0x2000</string>
  <key>Timeout</key>
  <string>2</string>
  <key>Legacy Logo</key>
  <string>Yes</string>
  <key>PCIRootUID</key>
  <string>1</string>
  <key>EthernetBuiltIn</key>
  <string>Yes</string>
  <key>UseKernelCache</key>
  <string>Yes</string>
  <key>GraphicsEnabler</key>
  <string></string>
  <key>Graphics Mode</key>
  <string></string>
  <key>device-properties</key>
  <string>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</string>
</dict>
</plist>

Now the system boots happily and actually shows the 1GB of VRAM on the NVIDIA GTX 465.

Check it out...

Big source for material compiled above is this forum post

WTF Google?

2011-09-26T12:03:00.000-04:00

Can someone please explain to me why the first match for "edit freebsd password file" isn't something that says:

# vipw

Seriously Google, that's all I wanted.

OS X Lion with GTX 570 and GTX 465 quad monitor support

2011-09-11T13:18:00.001-04:00

Disclaimer: if you've stumbled across this article don't attempt these steps without reading through the whole post and know that if you do screw something up your system may not boot again properly until you put your drive into another Mac or Hackintosh and undo your changes.

When I first purchased my ultimate Hackintosh machine, I bought the EVGA NVIDIA GeForce GTX 570 card. I did so foolishly without first finding out whether it was "supported" by OS X or the Hacktinosh community at large. I just got it because it sounded awesome and like a typical Linux user, assumed the drivers will come later.

Before we get going with the solution to getting the GTX 570 card working, here's my system specs:

Asus P6X58D-E Motherboard
Intel i7 970 processor 3.2GHz 6 Cores, 12 threads
6 - 4GB Kingston DDR3 1333MHz - 24GB RAM
2 - 2TB Western Digital Black SATA3 6GB/s drives

After a few attempts I was able to get the GTX 570 to boot and after a few more I was able to get it to boot up with correct resolution detection, it wasn't until 24 hours later that I was able to get everything to boot up with OpenGL (which was the biggest headache).

Okay so here's what I had to do to get it working... I ran into countless dead ends with OS X 10.7.0 so, refusing to be beat, I updated to 10.7.1 and that's when everything started working correctly for me.

I'm running:

Mac OS X Lion 10.7.1 Update
Chimera 1.5.4 (which added a device ID for the GTX 570)

Another important note with my rig is that my EVGA NVIDIA GTX 570 has the device ID 1081. There are others which have 1086 instead and so they need different configuration modifications than those listed below. (Possibly just changing 1081 to 1086 wherever it's listed below)

First thing's first, once you have OS X 10.7.1 installed, you need to edit the /System/Library/Extensions/NVDAGF100Hal.kext/Contents/Info.plist to add the device ID for your video card. I like to do stuff right from the Terminal though there are countless helper apps, so first let's create a Backup folder for all of the files we're going to break in case we need the originals again...

$ sudo su
Password:
# mkdir /Users/<username>/Desktop/Backup-GTX570-Mods/
# rsync -av /System/Library/Extensions/NVDAGF100Hal.kext/ /Users/<username>/Desktop/Backup-GTX570-Mods/NVDAGF100Hal.kext/

Now let's add our device IDs..

# nano -wT2 /System/Library/Extensions/NVDAGF100Hal.kext/Contents/Info.plist

Add your device ID to the IOPCIPrimaryMatch key list, mine is 1081 but yours may be 1086. The 10de part is nVidia's vendor ID, in case you're wondering. That's why they're all the same.

With this addition your system should boot up and you should have the correct resolution for your monitor instead of being stuck at 1024x768. YouTube videos will play but you won't be able to use any OpenGL or OpenCL like watching a video with VLC or opening Chess.

Next thing I added was ATY_init.kext which set things in motion. Once downloaded and unzipped...

# mv /Users/<username>/Downloads/ATY_Init.kext /System/Library/Extensions/
# chown -R 0:0 /System/Library/Extensions/ATY_Init.kext
# chmod -R 755 /System/Library/Extensions/ATY_Init.kext

There are three more files I needed to modify to get everything swimming along, the first was /Extra/com.apple.Boot.plist or, in my case since I'm running Chimera 1.5.4 /Extra/org.chameleon.Boot.plist In there I needed to disable the Graphics Mode by setting it to a blank string and add a Hex EFI String for the GTX 570 card. My Kernel Flags section also needed to be modified as it had something about i386 in it. Here's my /Extra/org.chameleon.Boot.plist (update sept 16 at bottom of this post has new file)

<plist version="1.0">
<dict>
  <key>Kernel</key>
  <string>mach_kernel</string>
  <key>Kernel Flags</key>
  <string>npci=0x2000</string>
  <key>Timeout</key>
  <string>2</string>
  <key>Legacy Logo</key>
  <string>Yes</string>
  <key>EthernetBuiltIn</key>
  <string>Yes</string>
  <key>Graphics Mode</key>
  <string></string>
  <key>device-properties</key>
  <string>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</string>
</dict>
</plist>

To generate my HEX string I used OSX86Tools. Open it up (reboot may be required?) and press Add EFI Strings/Boot Flag in the bottom box. Click on the GFX Strings button in the dialog box that pops up and then for Graphics Card choose Custom GeForce... at the bottom. Enter: NVIDIA GeForce GTX 570 for the name, Custom... for the VRAM and enter: 0x50000000 which equates to 1280MB. Select DVI-I / DVI-I and continue.

I just used the HEX value generated in the bottom window and created my new device-properties key manually. I didn't use the program to manipulate the org.chameleon.Boot.plist file, I did it via Terminal manually.

So now the system will boot and not set the graphics mode and it will be told that it's OpenGL capable. The next file I changed was a hex change from netkas to enable OpenCL for the card. So let's back up the files and make them user-editable.

# rsync -av /System/Library/Extensions/GeForceGLDriver.bundle/ /Users/<username>/Desktop/Backup-GTX570-Mods/GeForceGLDriver.bundle/
# cp /Users/<username>/Desktop/Backup-GTX570-Mods/GeForceGLDriver.bundle/Contents/MacOS/GeForceGLDriver /Users/<username>/Desktop/
# chown <username> /Users/<username>/Desktop/GeForceGLDriver

Now let's use HexEdit and follow netkas' instructions on our /Users/<username>/Desktop/GeForceGLDriver file.

Find
a883f8028b45f0
replace 02 with 03 to get
a883f8038b45f0

Find
ea83f802498b
replace 02 with 03 to get
ea83f803498b

So with our Desktop copy of the file patched and our original safely backed up let's put this one into place...

# mv /Users/<username>/Desktop/GeForceGLDriver /System/Library/Extensions/GeForceGLDriver.bundle/Contents/MacOS/
# chown -R 0:0 /System/Library/Extensions/GeForceGLDriver.bundle/
# chmod -R 755 /System/Library/Extensions/GeForceGLDriver.bundle/

So now there's just one more file left to tweak, it's the AppleGraphicsPowerManagement or AGPM. For our changes to take effect the system type has to match our defined system type and our GTX 570's device ID. So I used Multibeast from tonymacx86 to change my system type to iMac12,2 because that's what others had done and that's where the changes in this file seem to be.

# rsync -av /System/Library/Extensions/AppleGraphicsPowerManagement.kext/ /Users/<username>/Desktop/Backup-GTX570/AppleGraphicsPowerManagement.kext/

Open up /System/Library/Extensions/AppleGraphicsPowerManagement.kext/Contents/Info.plist and edit the section which begins with iMac12,2 to be as follows:

        <key>iMac12,2</key>
        <dict>
          <key>Vendor10deDevice1081</key>
          <dict>
          <key>Heuristic</key>
          <dict>
            <key>ID</key>
            <integer>0</integer>
            <key>IdleInterval</key>
            <integer>250</integer>
            <key>SensorOption</key>
            <integer>1</integer>
            <key>SensorSampleRate</key>
            <integer>4</integer>
            <key>TargetCount</key>
            <integer>5</integer>
            <key>Threshold_High</key>
            <array>
              <integer>93</integer>
              <integer>93</integer>
              <integer>93</integer>
              <integer>100</integer>
            </array>
            <key>Threshold_Low</key>
            <array>
              <integer>0</integer>
              <integer>97</integer>
              <integer>97</integer>
              <integer>100</integer>
            </array>
          </dict>
          <key>LogControl</key>
          <integer>1</integer>
          <key>control-id</key>
          <integer>18</integer>
        </dict>
        <key>default-control-id</key>
        <integer>18</integer>
      </dict>

So! After making all of those changes and rebooting, I was succesfully able to boot, MUCH faster than ever before and once the system booted up OpenGL/CI worked videos play in VLC, youtubes tube, and 3D benchmarks in NovaBench and Xbench don't crash the system anymore.

I tossed the EVGA nVidia GeForce GTX 465 in the second PCI Express slot and without any futher alterations I have quad monitors with 1920x1080 each. There's a slight flicker when I move a playing video from one monitor on one card to another but I seem to be able to play them halfway between each just fine.

A big thanks goes to the materials found from these threads on tonymacx86's forum and InsanelyMac's. Wouldn't have been possible for me to figure this out without all of their leg work.

Update [sept 13]: after only two days of use I noticed that my sound card wasn't working. I used Multibeast to add in the missing kext files and I've been working on recovering from that ever since. I can boot the system fine, but once again no OpenGL. In the process of trying to fix things I've updated to the new Chameleon layout which got rid of /Extras/Extensions and added UseKernelCache key. If I don't boot using the kernel cache, I can't boot. If I do use the kernel cache, it ignores my Graphics Mode declaration (I'm guessing that's why the OpenGL isn't working). Trial and error is becoming extremely annoying, but I will solve this... and then never update my system again.

Update [sept 16]: after a week more of trial and error, I've got everything working correctly again, I've had to alter my org.chameleon.Boot.plist because with the new Chimera versions it seems the Graphics Mode function is ignored, BUT, I think the only thing we need to avoid is declaring the graphics mode to the card before the OS boots (so the OpenGL) will work. So to do so we can just boot with -v and the graphics mode will only get set once the OS boots. Here's my new /Extra/org.chameleon.Boot.plist

Check out the write-up on 10.7.2 for more.

Using the OS X built-in tftp server

2011-07-27T15:18:00.002-04:00

I needed to configure some devices on the go and as long as you have a Mac handy you can do it without any additional software. I found details on another site but the /sbin/service they used didn't seem to exist on my machine but after a little more Googling I found this is the command to start the tftp server on OS X 10.7 Lion...

sudo launchctl load -F /System/Library/LaunchDaemons/tftp.plist
sudo launchctl start com.apple.tftpd

The default tftp file path is /private/tftpboot but you can change that by editing /System/Library/LaunchDaemons/tftp.plist and re-starting with launchctl...

sudo launchctl restart com.apple.tftpd

Characters skipping in nano when using ssh on OS X Lion 10.7

2011-07-22T18:20:00.002-04:00

I noticed using Terminal to connect via SSH to an IPv6 server and opening a file with nano, character repeats and skipping of the display when pressing and holding backspace or when typing rapidly. This is only new since I installed OS X 10.7 Lion. I seem to have solved the issue by going into Preferences for Terminal.app and selecting the Advanced tab and changing Declare terminal as: to xterm-color from xterm256-color.

Hopefully this helps someone else.

APN Issues While Travelling

2011-05-23T11:34:00.006-04:00

It is dead simple to jailbreak your iPhone, which allows you to install whatever you like and if you haven't updated your phone with iTunes ever, you may also be able to unlock your iPhone as well which allows you to use it with any carrier.

Recently while on vacation in France I purchased a full sized SIM card from Bouygues Telecom and that worked well for me but their pay as you go plan drained my 25 euro credit within a day in my old iPhone 3G because we were using Google Maps to navigate. So by the time I trimmed down the SIM card to fit as a MicroSIM for my iPhone 4 it was useless.

Since the credit was all used up and their data was far to expensive, I wasn't attached to them and found Orange France I jumped ship. For roughly 8-10 euro you get a SIM or MicroSIM with 5 euro worth of credit on it. You can add more as you see fit. The next step of the process is once Orange verifies you're human (roughly 24-72 hours after you sign up) you can set up the Internet Max plan for 1 month of Internet access for 9 euro. That plan gives you "unlimited" internet up to 500MB which for our two-week visit is more than enough, even if I want to tether.

So, I loaded up my SIM with 25 euro, plus the 5 that came with the card and a bonus 5 for buying 5 so I had 35 euro to last me until I was identified as a human in their system. During this time my internet should have been working an apparently it was sending requests out because my completely depleted again within a day, even though I never successfully viewed a map/tweet/site/email on the device.

I followed the instructions the helpful Orange attendant wrote down for me to register for the Internet Max plan roughly 24 hours after I bought the credit (the last steps of the process won't work if you're not "a human" yet").

Call #123#
Reply with 2 (Menu) Reply
4 (vos bon plans) Reply
4 (votre multimedia) Reply
6 (Suite [Next]) Reply
3 (option Internet Max) Reply
1 (Suite)
1 (Souscrire)
1 (Valider)
I can't remember what happened after that but basically, just continue until you're done. At this point you'll either have the plan added or it will say it can't yet.

Once I was pretty sure my Internet Max plan was added, my phone still wasn't working so I knew from previous travels that my problem was the APN settings. Access Point Names on cellular phones tell the phone where to direct traffic to connect to the internet. By default they're provided by your provider to Apple so when you plug in your phone, it just works. If you're jumping on a different network however, you may need to change the APN, as was the case for me.

To change your APN go to http://unlockit.co.nz on your iPhone, using a WiFi connection. Select Continue then Change APN and put in your APN settings. Lots of Googling may be required to find the settings for your provider. For Orange France the APN settings are:
APN: orange
User name: orange
Password: orange
(I'm confirming that as of now, May 2011 that works, so does using: orangenew.fr as the APN)

Once the Unlockit.co.nz profile is installed on your phone, reboot it. Once rebooted try to access a web site and if you're still getting errors about being "not subscribed to a cellular data package" try ejecting the SIM card tray and then re-inserting, without rebooting the phone. That was the clincher for me, once I did that it took seconds to pull up a site and we were on our way!

OS X Backup with rsync

2010-10-05T14:13:00.002-04:00

The site in which I found this has since disappeared, so for my own sanity, this rsync command line is a nice way to backup your currently booted Mac...

rsync -xrlptgoEv --progress --delete / /Volumes/NewDrive/

I'm going to blame Google Instant for the fact that I can't find the original page it came from anymore.

Old Forum Posts To The Rescue!

2010-09-01T11:10:00.007-04:00

I can't count the amount of times Google's answered my questions by mining old forum posts and solved my problem. Today yet another one...

One of my scripts needs to take ownership of a file to get access to it so I have the apache user set up in /etc/sudoers to be able to use chown with no password so that it can take ownership of the file and do what it needs to do. The problem was when I switched servers the new server being CentOS has a line in the sudoers file which makes sudo only work on a terminal session.
Defaults requiretty

Unfortunately if your command is run using PHP's exec() function then you don't get a response back from sudo because it comes back as an error so it's not until you run the function with the error catching on...
<?php exec("/usr/sbin/sudo /bin/chown username /file/to/own 2>&1", $output); ?>
that the output will actually have a helpful error message in it telling you just that; "sudo: sorry, you must have a tty to run sudo" which obviously can't happen from the web without removing the above offending line from /etc/sudoers

Here's my helpful forum post from 2007

iPad Developers are Greedy

2010-07-12T19:37:00.009-04:00

I've been working on the iPad for a week or so now and there's one thing that's become abundantly clear, there's an App for that but it costs five times as much as it does for the iPhone. The Apps on the iPad are not much more complex than their iPhone counterparts but for some reason the developers have all learned from each other and have all decided that instead of the go-to price of 99 cents they're all going to charge $4.99 for the exact same app on a bigger screen. I for one think it's ridiculous.

I agree that programmers require compensation for their work, but when your amount of work is nearly identical (if not halved by being able to use the bulk of the same code) on a big screen as it is on a little one, I don't see how you can justify the price difference.

Personally, I have no intention of buying an app unless I know it's going to suit my needs and work the way I want it to. The only way that is going to happen is if you give me a "lite" version of your apps to try before I buy or you give me the whole thing for free.

Hopefully the iPad situation only exists because developers were waiting for the iAd platform and we'll start to see some more free apps that are ad supported in the near future.

I think Apple missed the boat, they should have app rental. Try the app for free for a week and then if you want it, buy it.

Things I need: something to read Google Reader, as their web app sucks monkey nuts on the iPad. Reeder looks cool, but I'm not paying for it until I know it does what I want it to or you tell me it syncs my read list so I don't need to do things in more than one place.

Also, all of you people who are still serving up videos with Flash, get with the times, use HTML5's video tag and fall back to Flash for those pleebs that don't support the new era (aka: Microsoft Exploders).

Where's my iPhone 4?

2010-06-24T08:21:00.005-04:00

Okay Rogers, Bell & Telus. The Americans have their iPhone 4 in hand and operating. We have iOS 4 installed on our 3G and 3GSes so it's now your turn to, at the very least, give us a product announcement, a launch date, something!

FaceTime me.

iPhone 4 Announced

2010-06-07T21:14:00.004-04:00

Today Steve Jobs was at his best announcing Apple's newest iPhone, the iPhone 4. There were a lot of rumours as to what the new name would be but they've matched the iPhone 4 with their new operating system which runs on iPads & iPod Touches as well: iOS 4.

The big thing I was waiting for was the last announcement on the keynote. Video chat with a front facing camera? When I re-read that last line I thought to myself: why? For me the video's not a big deal, but with the video will come more chat applications and Apple's new FaceTime platform will run on open standards. That spells awesome interoperability options down the road, perhaps I'll finally be able to natively call my cell phone over WiFi using SIP without using an App that can only run in the foreground and eats through the non-swappable battery.

Also now that we'll have apps that can run in the background I'd love to see a Google Latitude-type application that can follow me around while I'm walking, riding, or running. Mix your GPS data with your photos and you've got awesome geotagging of your adventures around the globe, hopefully without ripping through batteries as well.

They boast the power sipping design of their A4 chip and their 300 hours of standby time is quite impressive considering I only get a day and half or so out of my iPhone 3G now that I've disabled all mail checking.

Apple seems to take their time getting things done because they like to release things right the first time. I respect that as I myself seem to work like that, but at the same time in our information age we want everything we want and we want it now. This product is no exception, give it here now.

Empire Avenue

2010-06-03T15:14:00.005-04:00

I've been tinkering with this new site Empire Avenue which looks like it's going to be the next big thing in the web 2.0 world. Highly addictive online world of buying and selling shares in people and companies/web sites. It's ridiculously entertaining if you enjoy buying and selling your friends and family and their startups. You earn credits for everything you do on the site or the sites you're connected to and it builds credit for for everyone who has invested in you.

The creators of the site are from Edmonton which is awesome because you've got to love supporting a Canadian site as a Canadian.

The site is currently in beta so you have to get an invitation into the mix but definitely worth checking out, just make sure you keep coming back or your stock will end up taking a dive. If you want an invite holler @ me: @ts2_ca

Import Garmin workouts to Nike+

2010-04-11T17:18:00.005-04:00

When I started long distance running, I bought a Nike+ sportband. I signed up for Nike+ and started clocking my running. Unfortunately the first revision of the sportband was so bad, it got recalled. I got myself the new (at the time) Garmin Forerunner 405 which not only tracked my mileage but also my position, even cooler. So for almost two years now I've been tracking my mileage and recording it on my PC. In that time the Garmin Connect web site has gone through a significant upgrade and now the web sharing is much more comprehensive and so has the Nike+ site. The only problem is that if you want to do both you have to run with two devices.

That was until Angus Smithson came along and created the Garmin Forerunner TCX file to Nike+ Converter & Uploader. Using his web site you can upload a TCX file from Garmin Training Center and then convert it and upload it to Nike+. Even better: once your run is on Garmin's Connect web site, you can just copy the activity ID and have the workout imported into Nike+ without having to export and upload the tcx file!

All you need is to generate your Nike+ PIN which never changes by following the simple instructions on his site and you're off to the races. Now you can compete with other Nike+ users in challenges and you don't have to weigh yourself down with an extra piece of equipment on your run.

I imported a lot of my runs from last year but unfortunately I hadn't synced my 405 in a while and I lost GPS points from some of my older workouts, the only one of which I found disappointing was my Dublin Marathon track. Lesson learned: sync up more often than every 4 months if you want all of your data.

Update: I found my backup and I had synchronized my Dublin Marathon track so I was able to add it to my Nike+ profile too!

My Nike+ screen name is Weezey1, feel free to add me as a friend!

OSX Update Fixed One Magic Mouse Issue

2010-04-08T10:48:00.002-04:00

When I leave my computer at the end of the day I lock it by switching to the Login Window and when I came back the following day, most of the time the right click on my Magic Mouse wouldn't work anymore. Only by rebooting could I bring that right click back on board. Now that I've updated to 10.6.3 however that right click works just fine all of the time.

Unfortunately now however after leaving my computer at the Login Window for more than a day (ie: weekend or work from home for a day) when I return my Magic Mouse's right click works, but the scrolling no longer functions. Once again, I must reboot my PC.

I guess I'll have to wait for 10.6.4 for the mouse to work through down time, or perhaps create a mechanical rodent that will move my mouse for me every hour.

Found the Missing Link

2010-03-29T10:03:00.003-04:00

While playing with my Hackintosh today, I fixed one of the things that aggravates me fairly often. I'm not sure why they decided it's a feature, but OSX has the option of skipping over select boxes and submit buttons when you press the tab key in your browser. I was looking for a Firefox fix and apparently there used to be one but now it's built into the operating system, which is why it took me longer than expected to figure it out.

To disable this "feature" open up System Preferences -> Keyboard and then under the Keyboard Shortcuts tab, at the bottom of the window select "All Controls" under Full Keyboard Access. Immediately your life will become better as you'll be able to press tab and up/down to select a list item or tab then enter to submit a form again. Everything Zen.

Technical Flaw in Gmail Migration

2010-02-02T08:03:00.004-05:00

Since I purchased Google paid storage (200GB for $50/year), I migrated all of my mail out of my Google Apps account and back into my original Gmail account. I started to do this by copying and pasting folder using IMAP access via Thunderbird with the accounts opened. That worked pretty well except for the fact that it was slow and then I ran into bandwidth limitations of Gmail accounts so the process started taking a really long time.

To get around that I used the recommeneded method of Gmail account migration which is to set up the source account as a POP account from within the destination account. That worked okay for a while, downloading 200 messages at a time and then it stopped working for about a week and I more or less gave up.

Today I noticed that it started working correctly again so I guess Google fixed something, but I found a problem with the process. A while back my email campaign system didn't specify the date in the header when it was sending out messages so there were lots of messages in my Apps email which didn't have date headers in them so the date of the message was the date received by the email application. So it seems now that my Gmail account has downloaded all of those Apps messages, the date specified on them now is when they were downloaded again, instead of the original date of download. This makes complete sense to me but it was rather confusing when I searched for some mail this morning.

The end result here is that my mail is now all in one Gmail account. 6GB of mail used of my 207.2GB of storage.

The iPad

2010-01-28T13:20:00.005-05:00

I don't think Apple has taken this much heat over a device that I can recall. People think it's cool, but fairly expensive and it doesn't allow you to run more than one program at a time. The biggest heat has come from the name. People are already referring to the largest fully loaded iPad as the Max-iPad.

The writers at MadTV had already built their own iPads back in the day.

Update: As expected, Hitler's reaction to the iPad:

LED Board Demo'd

2010-01-28T11:29:00.012-05:00

Last night I was asked to demo the board for Guelph's technical community. Other than the fact that it kept freezing it went very well. Check out these photos from @kevinleehansen.

Me being introduced with the questions; Who are you? What are you demoing? What do you hope to get out of it? and What should the audience get out of it?

Sign froze while I spoke, but it did get stuck on the world "Quality." Couldn't have planned that if I tried.

About the freezing: It seems to be interference that causes the sign to freeze. The program "making" the pixels continues to run. I actually thought of last night, resetting the COM port before each cycle that way if there was lots of power/data interference around it would be a little bit more robust and should reset on it's own. Also instead of the cables all piled into the box below a little more organization would probably help. (Keeping the power cables away from the TWI wires.)

Thanks to all for a great night and thanks to the roundup for the photos.

The Final Product

2010-01-21T17:12:00.004-05:00

Here's the end result of all of my hard work building the LED board.

SB-900 and Gary Fong Universal Diffuser

2010-01-01T21:01:00.002-05:00

I mulled over my dilemma and settled on the Nikon SB900 Speedlight.

This photo was taken in the evening, but it looks like a bright, sunny afternoon in my living room.

My camera gear is growing nicely. I'm going to need a camera bag. Maybe for my birthday I'll ask for that, and maybe one of these bad boys.

Gear Dilemma

2009-12-27T17:13:00.003-05:00

A friend of mine is looking to buy the same Nikon D90 I have, kit lens (18-105mm) and all. She has expressed interest in buying my gear as I'd love to upgrade.

My dilemma is that I would love to upgrade my camera to the new D300s. It's the next in the line of the Nikon cameras and has some damn awesome new features. Like the ability to write raw and jpeg images to a compact flash and SD card at the same time. Using that, I could use the Eye-Fi card I'm getting from Google because I upgraded my storage to 200GB for $50, to write my jpegs to that SD card and still write raw images to the CF card. Quality and awesomeness in one.

The other reason it's a dilemma is that I'll have to get a new lens if I sell the 18-105mm but I've been eye-ing a 50mm or 35mm prime for a while (drooling over the low f-stop) so it's not a simple purchase of the D300S body (which doesn't sell as a kit).

The final edge of my dilemma is that the actual upgrade I wanted to take on this season was a top of the line Nikon SB-900 Speedlight. It re-cycles fast, it only needs AA 4 batteries and with a Fong Diffuser produces bright, soft, shadowless light.

I guess I have to keep mulling it over.

Multiple Peggy2s showing serial video

2009-12-18T11:32:00.008-05:00

So I've been working in every spare second for the last 48 hours to get my LED sign board built and part of the project was to have it showing live Twitter feeds. Standing on the shoulders of Mr. Clegg and Mr. Oskay, this proved to be a relatively simple task.

I built my first Peggy 2 in about 5 hours and then started fiddling to get the Processing video working and within no time I could write static messages on the board. The next setup was to make it work with two Peggy2s. Since I only had a few LEDs left over from the first, but a about 75 left over from making these for our wedding, I recycled them into the second Peggy so I could start coding and see some kind of result if my code worked successfully.

I wired up the two peggys in paralell via a breadboard to the Arduino and started modifying code.

The code changes were actually quite simple, I made a second header and data array so that the second Peggy2 has a different header that it acts on to change it's state and then coded that into the sketch for the second Peggy2. I looked at changing the TWI address first but I think the Arduino code would need to be modified for that and right now it just spews data without modification and I didn't want to mess with that yet.

I made the image 400x200 instead of square and then I made the renderToPeggy function loop through each array of 25x25 independently and send out each header and each chunk of data. That was all it took and now each Peggy2 acts on it's own data and shows it's own pixels though both are receiving data for the other. (This could probably be more efficient but I wanted it to work right away.)

I wrote a simple PHP script to gobble up tweets that I was interested in by following the API guide to search and dump them into a database. Then another PHP script to distribute un-seen tweets via a URL and then modified the Processing code to ask for a new tweet after one has been displayed. If there's nothing new the code just shows a random tweet from the last 10 in the database.

Next steps in the project are;

get the rest of the LEDs for the second Peggy2
put both Peggys in some kind of a case, possibly with a coloured insert in front of the LEDs
~~add the ability in the Processing code to show a logo or many logos between tweets~~