Sunday, October 25, 2009

Google Wave First Look and Tutorial

Well I recently received my Google Wave account, and have been busy having fun with it ever since. Since Google Wave is actually getting a lot of press (which is unusual for a very beta tech product) there is actually a rather large public community on Google Wave, busily testing all of the features. Since I did get my invite, I thought I would share with everyone a little bit of how Google Wave works.

First off, the UI looks quite beautiful, and has many options for reorganization to maximize the space you have available to you. This will come very much in handy later, as you will see. Shown below is the default view you will see when logging into Google Wave.

The far left panes shows your "Contacts" and "Navigation" pane.  Currently, the contacts listed in the pane are all "extension" contacts, more about that later.  The "Navigation" pane is rather self-explanatory, and is similar to the gmail navigation pane. The "Inbox" is somewhat self-explanatory, and works in a similar way to an email inbox, yet also adds the functionality of an RSS feed of sorts. Notice the number highlighted in green under the time for each wave.  This is the number of unread messages in each wave.  When viewing the wave you will see wavelets/blips/messages that will either have a green bar on the left hand side, or they will be encircled in green.  These are the unread messages.  Clicking on them will mark them as read.   The view to the far right is the "Message" view, which shows all of the wavelets and blips in the wave.  Essentially, a wavelet is a threaded conversation inside of a wave, and a blip is an individual "post".  For an in-depth terminology reference see here. The nice thing about the UI is its ability to be reorganized. For example, reading a very large, in-depth wave will become tiresome if you have to read it all in a tiny window on the right hand side. So, as shown below, you can minimize or hide windows.  Notice the Inbox, Navigation and Contacts windows are minimized at the top of Google Wave.

There are essentially two types of waves: public waves and private waves.  To see a list of public waves, go to your search bar and type with:public which will bring up a massive list of public waves.  At this current point in time, public waves are a massive mess, as there is no easy way to manage/garden them.  It is akin to a massive global chat with embedded images/gadgets/bots.  I believe the real potential for Wave is with private waves, which are waves that are invite only.  This allows you to invite only people you trust, so that you can be productive inside of a wave.  As well, as it currently stands, if you click on a public wave, it will automatically be joined to your inbox, and with a massive amount of people constantly adding posts, it becomes a rather large mess.
If you look at the image above, you will notice that there is a extension added to the wave called "ClarkPoint".  This is a very handy application, as it allows you and anyone you add to your wave join in an instant conference call.. using VOIP.  All each of you need to do is click on "Group Call" and you will instantly be in a conference call.  As an added benefit of being in a wave.. all of you will be able to add notes simultaneously, in real time.  A very big benefit indeed.
Next, lets cover extensions/bots/gadgets.
Gadgets are essentially applications that run inside of a wave.  For example, lets say you wanted to see if everyone in you wave wanted to go on a trip.  You could insert a "Yes/No/Maybe" gadget.  This would allow everyone in the wave to cast their vote.  You could also add a map gadget, and everyone could see you (in real time) zoom into and out of the map, and add markers, and they could do so as well (in real-time).
Bots are added as users to your contacts, and you can add them to your wave, and (as far as I know) affect the entire wave.  For example, you can add a bot called "" and all of your conversations (the ones that occur after you add the bot) will be converted to speech bubbles, as shown below.     
Another bot, called, integrates wolfram alpha (for more information on wolfram alpha click here) data into a Google Wave.  To use this functionality, do a wolfram alpha search surrounded by brackets "[]" and py-robot will automatically post directly under you with the results of your search... very impressive.  As it is still in beta, it does not have all of the functionality of wolfram alpha.
Wave Editing and Navigation
At this point in waves, everyone can edit anyone's blips.  Obviously, this is a serious problem, and hopefully one that will be addressed in a new version.  Essentially, to reply to a wave with a new post at the bottom of the wave, hit the reply button on the upper menu bar, shown below.  The next button, the playback button, will playback sequentially everything that happened on the wave.  The archive button will archive the wave thereby removing it from your inbox.  The mute button will also remove the wave from the inbox, but make it available for searching.  The spam, read, and unread buttons are all features of email, and operate the same way.  The trash can will trash the wave.  The move to button will move the wave to a folder of your choice.
 When editing a wave, if you want to reply directly to another post, you can click on the arrow at the far right to drop down options and select reply to this message.  Private reply allows you to reply only to the message's author.  Clicking "Reply to this message" will reply to that message, and not at the end of the wave. 
Inbox and Wave Management Theories
I believe that the Wave "Inbox" will be completely unlike an email "Inbox".  I think that the amount of information that Wave will generate will explode making the classic inbox model unstable.  I am positive that due the extreme flexibility of the Wave platform many websites will use that for their forum, in at least some capacity.  Obviously, if the forum is being used heavily, you do not want that wave popping to the top of your inbox every time someone changes or adds to the wave.  You would need a way to specify notifications, or at least have a way to filter the wave to get the information you need.  I believe the trend that will emerge, (and believe me, not a new trend really) will be a search based inbox.  For example, you would create custom queries for different inbox views and switch through those views as needed. Some examples are with:me and with:public, for a full list, see here.
Currently, public waves are massive beasts that have you scrolling through hundreds of posts, and it is possible for dozens of those posts all throughout the wave to be edited at the same time.  This forces you to scroll thorough the page each time to find the specific thread you were looking for, in a vastly changing, confusing mess.  There needs to be a way to at least jump directly from unread blip to blip.. history will technically perform this, but at this point in time is much to slow and unwieldy to use in this manner. Perhaps a tag-cloud bot could be written to jump to certain parts of the wave... I think that could be something very useful.
Another difficulty encountered is chat management.  I was testing out a conference call gadget, (add it to the wave and all of the members can click add to conversation and the flash based application will use the microphone and camera to connect all the users) and someone chatted to me asking where I found the gadget.  The conversation went as follows:

hey.. sry.. mike not on right now lol Where do I get this gadget? THanks and you just dropped itin hear? one sec.. let me look it up.. searched the web to find it. I found it. Thanks though. Click on teh Green button to add by URL.. should add it.. oh okay.. cool.. there are some other ones like it... they work well too.. no problem. yeah cya. cool see ya.

Obviously, we were both typing in the same space, deleting some of our own text, some of the other person's text, and something that may have made sense to us at the time is now a jumble of unusable text.  Using playback is not an option to figure this out, as it is not sensitive enough to catch all of our little edits.  This is something that needs to be worked out.
Overall, I am very excited for the potential that Google Wave holds. I'm looking forward to the community to develop Google Wave into a very valuable product.
[edit]  Now you can check out my YouTube video covering the basics of Google Wave!

Saturday, October 24, 2009

Exchange 2007 Installation on Windows Server 2008 (Part 1)

Recently, I decided to install Exchange 2007 on a VM for training purposes.  The setup was as follows:
1. Virtualbox - for virtualizing two machines.
2. Windows Server 2003 - AD/DNS and Exchange Client.
3. Windows Server 2008 - Exchange 2007
The networking for these machines actually was really difficult and involved much more work than should have.  My plan was to have the machined network be on its own internal subnet, yet have access to the internet via a bridged interface.  However, I was unable to easily get the Virtualbox host interface to easily give 192 addresses to the internal VM's while maintaining an external 10 dot IP address. In the end (against my better judgment) I gave the internal VM's a 10 dot ip address and attached another nick to the host machine and bridged that interface with the VM's.  Needless to say the VM's were completely updated and locked down as soon as possible. On my Server 2003 VM I ran dcpromo and configured AD and DNS. The domain I named "test.local", an oversight considering my eventual plan was for Exchange to accept and send mail to external addresses.  Next I installed Windows Server 2008. (Note: With Virtualbox and Exchange 2007 do not choose dynamically expanding storage. Exchange will complain about not having enough mailbox space.) Once Server 2008 was installed and joined to the domain, it was time to install the pre-requisites for Exchange.  The nice part about the Exchange install (and new Microsoft product installations in general) is that Microsoft has become "Information heavy".  Error messages have (in general)  become much more detailed and often contain directions and links to fixing the problem.  However the problem often arises that the messages are much more verbose, and yet still as worthless. If you did not know the prereq's for an exchange install when you insert the install disc it has a list for you to follow, as shown below.

On a lot of servers, .Net framework and Microsoft Management Console will already be installed.  Next, Powershell will need to be installed.  Powershell is a very handy server management tool.  Windows is apparently going the way of linux for the ability to manage server without wasting the server's resources by forcing it to display a GUI.  

After installing Powershell you should be all set for starting the actual exchange install.  However, first, Exchange 2007 will test to make sure that your domain passes all of the requirements for installation. For more information follow Microsoft's requirement list, found here.  When the Exchange installer ran on my server, it discovered I had unmet requirements.  The requirements I needed were IIS.  I installed the following:
  • The default IIS 7 package
  • IIS 6 Managment Tools
  • Static and Dynamic Compression
  • Basic and Digest Authentication    
After installing the following components, the Exchange install started without any difficulties. The Exchange install took a rather large amount of time (over 40 minutes).  Everything appeared to install without incident.  After creating a snapshot of the hard drive, I rebooted the server. Next, I ran the Exchange Management console. Upon loading, Exchange loads the Finalize Deployment Wizard.  The next recommended thing to do is to run the best practices wizard, to make sure everything is working correctly.  This option is found in the toolbox.  The wizard will initially look for updates before running, and once it is updated it will direct you to the welcome screen.  From the welcome screen it will have you select an AD server to connect to.  Considering I only had one, it was a rather easy choice.   It gives an estimated time of one minute, however in my case a few seconds was all that was needed. Next you come to the main page, which has several options.  First you can select the scope of the scan... for me I obviously chose the only Exchange server I had running.

Next, I could chose from several different types of scans.  I chose the Health Check, which I believe is something you should run the instant you add an Exchange server.  If this was a production environment, I would perform the Performance Baseline Check as well to have something to compare future performance against.

The Health Check took about 3 minutes to complete, and then gave me an option to view the report, which I did.  Immediately, the report showed any critical errors.  I had one critical error, namely "Offline address book definition is missing."  To fix the error, I performed the following.  I went to Server Configuration -> Mailbox -> Mailbox Database -> Properties -> Client Settings.  From there it will show that the "Offline Address Book" is missing.  Just click browse and select an offline address book.  
I am unsure why this error occurred, and why this was not able to install correctly without forcing me to correct it manually.
Now I had to add some mailboxes.  As far as I can tell in Exchange 2007, adding users in active directory does not add them to Exchange 2007, so I added them under Recipient Configuration -> Mailbox -> New Mailbox.  Adding them here would add them in Active Directory, and would give them an Exchange mailbox.  Quick and easy... I am not sure why adding them in Active Directory does not add them automatically to Exchange.  I believe this should be a feature.

Now, I was ready to begin testing.  The first thing I did was forward port 443 to my exchange machine, to allow external access to Outlook Web Access (OWA).  
I tested this from a remote machine, and this worked fine, I was able to successfully log in and send and receive emails. (Note: To run in "full blown" mode, OWA requires Internet Explorer.  Firefox (and I assume other browsers) can run OWA in "Lite" mode.  I recommend using IE, as the web interface is quite beautiful when using IE, as shown below.

One thing I noticed (and this I did not know about) was that if you have a Sharepoint server you can access the documents using the "Documents" tab in OWA.  Very handy feature, if my Sharepoint Server was up and running I would have tested that feature, as I'm curious as to how exactly that works.  Apparently you are also able to access Windows File Shares from that interface.
Next, I was able to install Outlook 2003 on my DC.  I started up Outlook 2003, specified that I was connecting to an Exchange server and everything set up correctly, and worked excellently right out of the box, as shown below. 

At this point in time I was quite satisfied that at least Exchange was working with minimal amount of work on my part.  Then, after getting Outlook 2003 working, I decided to move to Outlook 2007.  
I installed Outlook 2007 on the DC as well, and when I started the program, it found my user name in the Exchange Mailboxes, and automatically filled in my user name and password. I clicked next, and then it prompted me enter a password to connect to the Exchange Server.  However, no matter what combination of user names, passwords, and domains, I was unable to log on.  So I finally clicked "cancel", and then Outlook displayed the error "The connection to Microsoft Exchange is unavailable. 

Outlook must be online and connected to complete this action."  After clicking "okay" another dialog box comes up.  That dialog box has you resolve the name for the server.

After making sure everything was correct, I clicked "Check Name", however Outlook was unable to resolve the name.  Nslookup queries proved successful, and considering that Outlook 2003 worked fine, this was an interesting error.
Next, I attempted to manually connect to the exchange server.  However, this also proved impossible.
My next post will cover everything that was attempted to resolve the problem, and more about Exchange Server 2007.

Wednesday, October 7, 2009

Switching from Hamachi to NeoRouter

Due to the problems I kept on having with my Windows 7 laptop Hamachi connecting to my Debian Server Hamachi, I have switched VPN solutions.  So far I have been pleasantly surprised... it has all of the same features of Hamachi (as far as I can tell) and a few more nice features of it's own.  For gaming, hamachi, unfortunately, was slightly weak.  Trying to play a CS:S game would fail, as the addresses would not all be class C address.  However, with NeoRouter, the IP's are all class C addresses (configurable), so this should not be a problem.

However, NeoRouter has a few more configuration steps that need to be taken for it to be set up, compared to Hamachi.  For one thing, you must install a server version of NeoRouter, in addition to all of the client versions.  Fortunately, NeoRouter has a Linux version in the form of a .deb package for easy installation on Debian.  Running the server install on Debian does not give you any configuration options... once it's installed it's installed and running.  Interestingly enough, you are unable to edit the server configuration at at all from a linux machine.. you must have a windows machine... in the same LAN (according to their documentation... you should be able to do it externally.. just have to prep your firewall first).  However, once you open the server configuration (shown below), you can edit the options.  The small screen on the right is the client view.  To access the server options, you need to go to File -> Options and then log on to server.  This is done by entering the username and password that you use to log onto your linux box.  The logon-to value is the IP address of the linux box.  That will bring up the screen shown below.  The logon-to value is the name of the "Network" that you created (The first time you logon this will be the IP address of the linux box).  The IP address box will show the IP address of the server, and below the port number (32976 is default).

Next, you are able to add user accounts (as shown below).  These are user accounts that your friends (or you) can use to log on.  The value that these accounts add is that you can assign "User" vs "Administrator" accounts, and give your friends user account credentials to sign onto your network.

The computers tab (shown below) helps regulate access permissions that users have to certain computers.  You can add users and then turn the firewall on for that user, and add or delete access to certain parts/ports on your computers... very nice, and it has a granularity that Hamachi does not.

The connections tab allows you to set P2P connection settings.  This is a very useful feature, as you can tell the computers in the same network to create direct UDP connections.  This should allow for faster connection speeds, as you are able to go directly from peer to peer without interacting with the server. As well, (this is a guess) you may be able to continue utilizing NeoRouter even if the server goes down (as long as the UDP connection stays up, once the UDP connection is lost, you will need the server).

Lastly is the settings dialog (shown below).  The "Current domain" shows what domain you are currently logged onto. This is the setting that you defined when you logged on by what you entered into the "log into" box.  When you first configure the server, you enter the IP address of the server, and you configure and join the network that way.

 However, once you leave the LAN, you would have to enter your external IP address.  This obviously becomes very difficult if you have a dynamic IP address.. you would first have to find your IP address before you entered it into the "log-on" box.  As well, every time that IP address changed you would have to re-enter it.  The domain option changes all this.  When you enter a domain name your NeoRouter client sends an https request to the NeoRouter servers (not your server) and checks to see what IP is registered to that domain... so essentially NeoRouter acts as a DDNS server for your VPN.  This is one step removed from Hamachi, which also sets up the connections... with NeoRouter your server still handles connections, but your clients contact NeoRouter to find the server.  However, I actually have a DDNS account, so I can enter my domain name and I will always be able to find my dynamic IP.

Now I'm going to explain exactly how this works, to the best of my ability.. hopefully this is accurate.  The server will receive all requests to join your "domain", (sometimes using NeoRouter's DDNS service) and will then (by default) connect each client directly using UDP, after assigning them all IP addresses.  The clients are then able to connect directly as a VPN.  If a client loses connection it will again contact the server.

The only other thing left to test is NeoRouter's ability as a service.  With Hamachi, you did not have to be logged on to Windows to use Hamachi, we will see if NeoRouter is the same way.

Hope that helps some of you if you wanted a better choice than Hamachi.

Thursday, October 1, 2009

Windows XP to Debian Server Switch

Recently, I decided to switch from Windows XP to Debian, and further more decided to go cutting (read: bleeding) edge and install Debian Squeeze (Debian 6), which is still in the "unstable" phase. At the bottom of this post, I have a Youtube link to a video detailing the installation.

The list of programs that I had on my windows "server" was quite normal, nothing extremely special:
1. Subsonic - for streaming music\video (with my internet connection the video really wasn't practical.

2. uTorrent - I'm kind of a packrat, and I have almost all of the newest versions of software in RSS feeds (Check out SARDU antivirus, UBCD, etc.. I have RSS feeds for all that software so I can download the newest version automatically). As well, I for some reason need to have the newest releases for my favorite linux distro's.. you guessed it, RSS feeds.

3. Hamachi - For VPN, because I need a better router before I will implement OpenVPN, don't have enough NVRAM in my router for it to work. Feel free to donate to help me buy a better router :D.

4. Apache - hosts my web server, which hosts my freemind maps, and other test environments.

5. Freemind - For creating the mindmaps... I had this application shared through another computer, however since I was redoing that computer I was hoping to find a way to remotely share applications.

6. Remote Desktop - As I am not linux-fluent, I would need the gui for somethings, as I would not be able to accomplish everything from the command line.

The following programs were eventually used on my Debian server:
1. Subsonic - since this is Java based it works on pretty much anything. The first install went without any errors, but did not appear to work. Purging the installation and reinstalling worked... not sure what went wrong. As well, I had to change the ownership of some of the files, to root ownership.. my default logon was throwing up permission errors. However, once that was corrected, subsonic was up and running very well. The next thing I had to do was get it to start on boot. By adding the following line to /etc/rc.local "/var/subsonic/standalone/" Subsonic will boot on startup... rather easy compared to some programs rather long startup scripts.

2. Deluge - Bittorrent was eventually taken over by Deluge. This was a rather difficult one... I originally went for Transmission.. I then realized I wanted something with a Webui. Transmission was said to work with something called Clutch. However, I was unable to get that working. From there I turned to Deluge.. however, the debian repositories appear to be very out of date, they only had Deluge .5. I downloaded Deluge version 1.1.9, and after MANY struggles with dependencies for compiling and installing I finally got it up and running. The only complaint I have is that RSS feeds cannot be managed from the Webui.

3. Hamachi - Hamachi was rather easily installed, however there is one error that occurs when starting.. for some reason the tuncfg command is not run... I believe this is a permissions issue, and is something that needs to be changed with my startup script. As well, there seems to be problems connecting my windows 7 laptop to my server using hamachi; the hamachi connectivity is sporadic. I am currently looking into another solution.

4. Apache2 - Apache has been replace by Apache2. The installation was once again, very easy on debian, sudo apt-get install Apache2. There has been a change with apache2. You now edit the apache2.conf file instead of the httpd.conf file. It is my understanding that you can add lines to the (empty) httpd.conf file and they will be applied. All I did was copy my backed up web files to /var/www and replace the index.html, and I was once again up and running.

5. Freemind - Freemind was easily installed on debian, and worked exactly as it did on windows.

6. NxServer - I originally tried VNC viewer, but VNC was unable to run the server headless.. or at least I was unable to get it working. So I installed nxserver, and installed nxclient for viewing. NOTE: Once I installed xfce4 for the desktop manager, then I was able to use X11 forwarding to forward the desktop to the windows machine.. more on this in another post.

1. Webmin - Webmin is an amazing web-based administration suite for linux. A simple install and you can connect via port 10000. Webmin has many advanced options for managing your server, from apache2 configurations, to Samba configurations, all on a nice web-based interface, as shown below.

Essentially, I was able to replace my windows server with alternative software.. now it was much more difficult to get up and running, since I was quite inexperienced with Linux, but now that it is up and running, it runs much quicker than my windows box was running.. although most of that is due to the fact that I am running the server "headless". Below is a video detailing some of the installation.

At this point, the server is up and running... I have not run into any unexpected problems (once I got everything running strongly, when first configuring I ran into a lot) but there are a few kinks to work out still, especially with X11 forwarding.