It was pretty straightforward, and I found a post that does an excellent job of summarizing it at http://fedorasolved.org/video-solutions/nvidia-yum-kmod

I will comment on one thing though – when I did the install, I used the basic kernel for my i686 32-bit CPU. I found that my 4 GB of RAM only showed up in the system information page as 2.7 GB! I looked into this a bit, and it turns out that if you want greater than 3GB of memory supported, you need to have a 64-bit machine, OR you can handle this in software with the PAE kernel.

PAE stands for “Physical Address Extension” and all you have to do in install the PAE kernel, and then also the PAE Nvidia modules, and everything will work fine!

I know have my machine up with 8GB of RAM and working perfect! I read that the PAE kernel will cause a performance hit (sort of like doing RAID via software versus hardware I suppose), but I don’t see anything like that on my system.

My new server has a nice Coolermaster case that can hold 10 drives, and has a 1100 Watt power supply. I really like the case, it makes putting in drives a snap, plus the cable wiring went very well. Plus the thing is made of cool brushed aluminum. Who can argue with that?

I was re-commissioning this server after having used it as a gaming Windows machine for a bit. So it had dual NVidia SLI cards (GeForce 8800 Ultra) in it. When I first booted up the LiveCD, it got kernel errors that didn’t really track down to anything that was helpful. It would boot up but then freeze after a few minutes.

I tried the DVD install, and that got the same results. After much head scratching and googling, I tried removing one of the Nvidia cards. I was planning on only using 2 monitors anyway, and each card had 2 DVI outputs. This did the trick!

Next up, I had to run keyboard/mouse cables to my home office. I like to keep the actual computer down in my server room in the basement, it keeps my office nice and quiet. But I found that my 100 foot run for my PS/2 mouse didn’t work! My mouse is a Logitech MX310 that is really USB but I use a PS/2 converter on it so I can use a long cable run. This had been working fine on my previous Linux installations.

I found that the mouse would work fine if directly attached, so it was a distance issue! Which is odd, since on my old server it worked fine, and the cable was the same. I ended up using an USB extender device that I had bought years ago and never needed. It uses ethernet cable in between two adapters so you can have a long USB run. This did the trick! The brand is “coolgear” and I forget where I got it.

Next up: Getting Nvidia working on Fedora 13

Now seriously, what the hell? They’re using a fading scale to indicate the uses? So I suppose “Multiple Computers” is vaguely recommended? And “Streaming Music” seems to be slightly recommended? I suppose the usage icon on the far right is downright not recommended, but can’t we just get an honest do or don’t from the usage guide?

And then the marketing team seems to have their own rating system which is not explained anywhere and is completely unhelpful. Ok, so “High Performance” is rated N++! Are you kidding me? What the hell is a “N”, and what does a plus sign indicate? What would “N–” mean?

The back of the box should be a case study in a marketing class on what not to do!

A quick tech tip… there are actually TWO WRT54G2 models floating around on shelves. DO NOT buy the one that has this silly thing on the back of the box, it actually has less onboard memory than the other model. But they both cost the same.

Here are the steps I need to take to fix them:

1) Press the power button to turn the unit off.
2) Disconnect the RF coax cable from the back of the unit.
3) Unplug the unit’s power cord from the wall.
4) Reconnect the RF coax cable to the back of the unit.
5) Wait at least one minute.
6) Plug the unit’s power cord back in and allow it to boot up.
7) Enjoy flawless ReplayTV operation.

I had problems getting the DVD bootup to go into graphical mode, it kept showing an error that it couldn’t start X, and would continue with text mode. Of course, this is less than ideal since the text mode doesn’t have all the nice install options (not sure why) so I needed to figure it out.

The easy solution was to remove one of the cards and try again! This did the trick, and now I have Fed11 nicely running on the one card powering two monitors. Another thing to note is that I had to install Nvidia’s linux drivers. Also note, that if you run “yum update” after installing those drivers, you will need to re-install them one more time since yum will overwrite their functions.

My next step is to plug in the 2nd card and see if it is recognized properly and I can get 3 monitors going in independent configurations…

[Update] I added the second card after the fact, and now have 3 monitors hooked up nicely with no problems!

The laptop has a little switch on the top of the keyboard that turns wireless on/off, and usually when you turn it off and back on, the computer will recognize the wifi hardware and it can find networks. But in this case, it wasn’t working, and no wifi networks were begin discovered.

After an hour of trying all sorts of things like checking out network settings, etc, it turns out it we had to use a function key (F1) to toggle between different Bluetooth/LAN/WAN settings! We had to choose the setting of Bluetooth and LAN to get this working. Apparently adding the Bluetooth device switched this setting, and the only way to get it back is to use this poorly labelled function key. Live and learn.

I had googled around to try and find help on this, but nothing pertinent showed up — so hopefully this post can help someone who has this problem in the future!

• a NFS file server for other UNIX systems
• a Samba file server for Windows systems
• a LDAP server for keeping a single location of contacts
• an Apache web server for an internal website of links and other information (photos, etc)
• a Slimserver daemon that provides support to my Squeezebox
• a MediaTomb server that provides support to my DLINK DSM-520
• an internal Domain Name Server (DNS)

and more!

I use this setup to hold all my digital photos, audio (mp3) files, and I’m now in the process of scanning in my DVD library. So let me take a moment to say “Just say NO to piracy!”

I personally have found it to be cost effective to buy used CDs or DVDs and scan them in to get good quality recordings, and all my media stays only on my local server that is inaccessible from the Internet. Aside from the legal issues, who wants to clog up their Internet connection with unnecessary traffic for others?

I’ll pass along some things I found while building my setup. I never went the iTunes route because I’ve always used mp3 players that were smaller and cheaper than the iPod. Plus, I want to use a generic format like mp3 and not be limited by what I can play it on. With mp3 I can play all my music on my computer, laptop, stereo (through Slimdevices Squeezebox)  or even my DSM-520. It is simple enough to get your CDs into mp3 files so I won’t go into that.

Next, the video portion. I chose the DSM-520 because I wanted a cheap box that had no moving parts (i.e., no hard drive) that would talk nicely to my fileserver. Well, it turns out that it doesn’t just look for a fileserver using something like Samba — it requires a UPnP type server to be running. DLINK gives you one, but it is for Windows, and you can’t really have a Windows box on all the time to that purpose. At least, not reliably Seriously, I have linux boxes that have been up for 450 days with no problems. I can’t let a Windows box go for 3 weeks before all hell breaks loose.

So, I looked around for some UPnP server software I could use and chose MediaTomb which I could easily add using Yum. It is simple enough to run, and lets me add video/photos/music that puts references of those files into a MediaTomb database.

Now you need to scan your DVDs into your library! There are various free packages (like handbrake) which I looked at, but in the end I settled on CloneDVDMobile which is by SlySoft, a great little company that I’ve used for their AnyDVD software which allows me to watch my DVDs on my PC (believe it or not, some DVDs refuse to play nice on a PC).

When you use this software, you are given a whole slew of formats to choose from. I used the generic DivX one that produces AVI files. Most DVDs will offer me the 852×480 resolution, and I max out the quality which goes to 25-29 usually. I use 2-pass deinterlacing and prefer saving it as one big file.

The AVI files it produces look great on my win media player, and I had some trouble figuring out how to make it look nice on my DSM-520, and after much hair pulling it turned out that I just had to alter the aspect ratio in the DSM-520’s settings screen to 16:9 and then it was perfect!

So now I can kick back on my couch and flip through all my DVD titles and watch anything I want with a touch of a button!

Of course now my wife says, “There’s nothing here to watch! It’s all The Matrix and James Bond!”

So now I need to scan in a bunch of Julia Roberts movies.

I assume you have already done the basics, like created your developer account (and know your secret key), and installed the command line tools.

Note that you will need Java installed on your system for the tools to function.

You can find the official EC2 docs here.

Now, on to my tips!

Find A Starter Image

You can browse the images at Amazon via the command :

You’ll see a bunch of images that are out there, in different UNIX flavors, and some with LAMP already installed. I found that whatever you pick, you’ll want to do your own package installs anyway, so just pick a baseline that you like, i.e. Ubuntu or Redhat or whichever you feel comfortable.

Note that you need the AMI string of your choice to continue. Not only will you see a number of Amazon created AMI’s, but other folks have created some public AMIs that you can choose from. Here are some AMIs for Ubuntu you can check out.

Choose A Server Size

Note that EC2 provides a small, large, and extra-large version your can use. I’ve found that the small instances are incompatible with the the large and XL, so be careful! You should start with the large if you plan on any expansion whatsoever.

So when you are ready to create an instance, you’ll need to specify the size (m1.large or m1.xlarge) on the command line otherwise it will default to small, i.e.:

You’ll need to create a gsg-keypair file first of course. This is described nicely in the command line docs, you’ll use “ec2-create-keypair”.

Note that you can apply a firewall to your instance via the “-g <name>” option, where you define a group with the name of “<name>”. You can see what groups you have via:

You can create these groups using commands like:

Access Your Instance

After you run an instance, you can check on the progress via:

And you can provide the instance name also, e.g.

Then you can SSH into it when it is ready by using the domain name it returns, e.g.

Customize Your Instance

You should now make sure your instance has all the packages you want, and configure them how you like. I install the latest Apache, MySQL, PHP, Perl, etc, and edit all the configuration files to my liking.fone

Make sure you stay in “/”, and do not use “/mnt” since this will go away if you reboot and won’t be saved when you save your instance.

If you need more space, you can use the extra storage service from Amazon although I have not done that at this point.

When you have everything the way you like it, you should save your instance. You will use ec2-bundle-vol to create the image, and ec2-upload-bundle to upload it to Amazon’s storage service.

You’ll need your secret key, access key, cert key, user id and a sample command set is as follows. You will run these from your Amazon instance (where real keys and numbers are replaced by X’s and fake numbers):

Once it is uploaded, you will need to now register it as an AMI image you can access later. You need to do this command from your local host:

This will output your AMI string that you need to note so you can use it later to use your new instance on new Amazon servers.

Cleanup When Done

Don’t forget to make sure you terminate instances you don’t need to use anymore, or you will continue to be charged for them! Run “ec2-describe-instances” to see what you are running, and then you can use “ec2-terminate-instances” with the instance string as an argument to remove them.