I just finished moving CornerPixel’s main website to ReliableSite.Net’s hosting.  Originally I had www.cornerpixel.net hosted on IXWeb Hosting.  The draw to IXWeb hosting was the unlimited disk space and unlimited SQL 2005 databases.  IXWeb hosting was good for a little while, and I was some what happy with their services and technical support.

Soon however things got slooooowwwwww.  I found that all that space was too good to be true.  Everyone else was jumping on the bandwagon and doing things like FTP was horribly painful.  My site being mainly just static content seemed to perform okay.  I also had several DotNetNuke sites on IXWeb hosting and they loaded fairly okay.  I was also having issues trying to install new DotNetNuke sites because the SQL server was so overloaded that the process would time out and fail.

So having said all that, I have moved www.cornerpixel.net to ReliableSite.Net.  The site is zippy and so far the support has been good. 

My next step is to move the blog.  I think I will use this opportunity to change from Das Blog to WordPress.

~Michael Silva - MikePixel

My last post was going over issues I had setting up NLB clustered servers within an ESX or VMware environment. 

I thought I had it all figured out and then I had another issue.  Whenever I would restart one of the NLB servers, the other server and the host would be unreachable.  I had a –t ping up to all three IP’s to troubleshoot the issue.

I did some research and found that VMware recommends to use Multicast, and that if you used Unicast, there were a couple steps to follow.   (Wouldn’t you know I was using Unicast.  We are using Unicast because the network guys didn’t want to add ARP entries to our switches at the moment.)

Being that the vSwitch sends out RARP broadcasts to the switch, you need to set the “Notify Switches” to be NO.

Below are the steps.

Within your VI Client

• click on the Host name in the left column.
• click on Network under “Hardware” column.
• click on Properties on the right (next to remove)

This window will pop up.

Next …

• click on VM Network as shown above
• click “Edit” button

This window will pop up.

Next ….

• click on the NIC Teaming tab
• make sure the “Notify Switches dropdown is set to No
• click OK
• click close on the vSwitch properties window

Now you should have NO issues. 

Remember however that it is not recommended to use Unicast with ESX NLB clusters.  Unless you run into some reason not to, I would recommend using Multicast.

I hope this helps,

Michael Silva - MikePixel

The last couple days I have been struggling a lot while setting up NLB (Network Load Balance) between two web servers.  I was beating my head up against the desk when I was trying to add the second host to the cluster inside the Network Load Balancing Manager on Windows Server 2003.

For background, I have a web server that we will call ws1.  It is working fine and is a guest on an ESX server.  I wanted to make two more servers out of this one so I used VM’s Converter.  This worked great.

Now I had ws2 and ws3.  The next step was to some how get the ESX Host to assign these VM’s with new UUID’s.  I ended up in the datastore renaming the folder with the VM’s in them.  What this does is it copies the entire folder, renames it and then deletes the old folder.  I think it does this because ESX is running on Linux server.

Once this is done, I can add the VM’s back into the inventory and it prompts me to do something with the UUID.

I chose to Create a new identifier.

Once this was done, I came across a different issue when I tried to create the NLB cluster.  I was able to create the cluster fine.

I also could add the first VM just fine.

But as soon as I tried to add ws3, I had an issue.  When I tried to connect to the host (ws3) I go the following error. “No interfaces are available for installing a new cluster”

What I found out was that their was an issue between the Network Connections.  Since they were actually both the same machine (since I converted both VM’s from the same server (ws1)), the Network Load Balance Manager was seeing the Network Connections as being the same.  In order to fix this I had to go into the VM settings for both ws2 and ws3 and add another network connection as seen below.

Once I started the VM up, I just changed the IP settings for the new Network Connection and named it External, and renamed the old Network Connection “Do-Not-USE”

After all that I could now add both servers to my Network Load Balance Manager and to the cluster.  I am not sure if this all makes sense, but for those who come across the same issue, I hope this will help. 

~Michael Silva - MikePixel

This sounds like a NO brainer but today I found out that someone had created an account and changed one of my skin pages.  They erased all the JavaScript I was using to allow users to change the background image.

I found out I hadn’t changed the configuration key for security/new-account-role. 

This could have been a BIG problem.  Hopefully my readers won’t have this issue.  To avoid this make sure to go to your Control Panel.  Click on the System Settings tab.

Once there Click on the Configuration sub menu and scroll down to the configuration key you see in the above picture.  (the one by the red arrow)  Change that to Viewer or whatever you want to change it to.

You can also keep people from registering by changing the security/allow-anon-account-creation to false.  You can also see this configuration key in the picture above.

Well I hope this helps.  It is something small, but could really get you in a bind REAL quick.  I think the default is set to Contributor, so PLEASE check this out right away.

~MikePixel (Michael Silva)