Well I found a program that seems to work for posting better then what I was using. Just if i buy it I will post more, hey you get what you pay for.Ecto seems to work better, it even lets me delete test posts. I should use an image as a test, but i will get to that later. Right now not a whole lot running around in my mind other then I have some work to do tonight at 1AM.

I think I might buy this app and find out how it posts things to get the picture plugin to work better. Time to start the coding I guess

I might poke at it later, but for now a simple weekly tip
When you need to check someone's mail server you can use nslookup on XP and Vista

nslookup
> set type=mx

> server 4.2.2.2
Default server: 4.2.2.2
Address: 4.2.2.2#53

> velcrohurts.com
  Server: 4.2.2.2
Address: 4.2.2.2#53

Non-authoritative answer:

velcrohurts.com mail exchanger = 10 mx.velcrohurts.com.cust.b.hostedemail.com.

Authoritative answers can be found from:
>

Now to review this. nslookup is the name server lookup for most OSes.

set type=mx
that tells nslookup that we want to look for a mail record or a Mail eXchange record

server 4.2.2.2
This chooses a server outside of most peoples controller. This allows us to see our information from an external source that is not in our control to make sure our records are correct

The next part is the domain that we want to look up. In my case i used my email server. By telling me that it is the record has not been changed for a while or updated.

If you want to get seriously into DNS, I recommend setting up a linux server like Ubuntu, CentOS, or Slackare and setting up BIND atleast once.

So I have been working with Exchange 2007 and it has a lot of advancements over previous versions. One thing that I feel it is missing is in the area of Unified Messaging. It can interact with a few PBXes and IP PBXs out of the box, and a few motivated people have been able to use sipX as a Sip Proxy to pass the traffic from an asterisk server to the Exchange server. It is actually really simple if you follow these instructions by Ryan Newington. It is a bit older, and some things you have to guess at because they have changed, such as trixbox configurations and layouts, but over all you can swap out trixbox for other asterisk distributions, just remember to RTFM.  It does help to have a familiar with Linux in general, there is a lot done on this with Red Hat distros, such as Fedora or CentOS. I had some interesting issues trying to get the trixbox virtual machine running, such as I had to re-install GRUB, and use VMWare Converter (thanks Kobold) to convert the machine image from an ESX image to a VMware Server image. At the moment though, my system is stuck at the authentication stage from one server to the sipX proxy. I can call from the trixbox to the exchange server, but I can not call from the outside to the exchange server. I will try to work this out tomorrow.

I am tired and starting to ramble so I will just leave it at this for tonight.

So the question is how to get a massive amount of users on to an email system that does not break.

The problems were

1. they already have existing data that they need to keep
2. they were on Qmail Toaster 
3. must have a web front end (for some users not all)

So in keeping with those requirements we decided to try to farm qmail-toaster.

The idea was to use a Foundry load balancer for the incoming mail connections that would delegate them out to the independent email servers.

The email servers themselves would have a shared back end mounted to the maildir, in this case /home/vpopmail and this would be done via a NetApp. Sadly the iSCSI implementation for Linux sucks so the only option would be NFS, on a private gigE network.

Also since qmail-toaster requires MySQL, the idea is to break that service as well. This would also make sure that when someone logs into any of the servers to get their email that they would be authenticated correctly.

By breaking some of the base services off to other servers it would make it so that everything would have shared resources across the farm. By breaking off the web services, we would be able to have two separate types of web mail, a new slicker interface, RoundCube, or the current qmail-toaster web interface. This makes sense when you take the thought that some people fear change and would complain if you changed their interface without warning.

  To the left is the mockup that was done for this layout. It should help with some of the basic information. The full image is here

I would like to state that you are on your own for trying to implement this. I do not claim anything here will work, it is just a conceptual mock-up to get a basic idea. We might implement this and we might not. It could fail miserably because of the spam filtering/anti-virus. There are many details to this that have not been touched on, so if you try it and it doesn't work do not blame me.

And the video