That’s pretty much it for today’s round of Dustin expo. Saturday I will be heading back digging deeper in to products land. You will find pictures on my Twitter: clicky

I’m eating lunch at the moment so I’m a bit short on time, I’ll make another post when the second speaking is done.

Peace!

In a traditional environment you would have an array of HDDs for every server acting as its main source of data. This could be both expensive and slow depending on how many disks you could afford to put in every server.

For example, a server that doesn’t need much storage could have problem running a redundant yet powerful solution like raid5 or raid10

This could look something like this:

(click to enlarge)

As you see here S2 has a lot of space left while S3 is almost full. This means you would have to purchase a new disk to extend the array on S3 while S2 had massive amounts of storage left over. Besides that, as mentioned before, there is usually the issue of performance and redundancy as well.
What you do in a SAN is that you combine your drives in a chassi and then connect all your servers via your choice of interface. The most popular ones (I think) are iSCSI and FC (Fiber Channel). We’ve chosen a Fibre Channel solution for our SAN which basically means that you connect your servers via fibre cable interface to the SAN. Since I’m very new to this I don’t know the exact good/bad side of this compared to iSCSI but our reseller recommended it so I’m going with the flow.

A san would look something like this:

(click to enlarge)

Pretty cool if you ask me! As you can see there’s now space to grow in and you don’t have to get new disks – Huzza! Besides that you can now run more complex/better redundancy/better performance types of raid solutions on _all_ your servers that can use a FC card.

All of the above are the good things, some bad things are

* Disk failures
Let’s say you have disk failure in your SAN. If you run raid5 (for example)  and more than one disk crashes you’re smoked on that entire array, this could span across several servers. You would then have to rely on backups to get your data back. In the “old” environment you would only have one server fail and it would be quicker to get back up running.

* Price
It’s quite pricy to get started, however I believe you do go plus with time depending on your expansion rate. Also, as some vendor in world of warcraft says… “Time is money, friend!”

If you have any questions or similar feel free to ask me in comment section!

Basically outlook anywhere is a way for users to connect to the Exchange server via HTTPS.

As shown here in this neat graph stolen from http://office.microsoft.com/en-us/outlook/hp101024441033.aspx

(click to enlarge)

As you can see it saves you the hassle of connecting to the business VPN every time you want to check your email.

If you’ve read my blog before you might recall reading I’m using Windows 2008 x64 with Exchange 2007 running on it.

There’s a big problem with this combination regarding outlook anywhere, and it has to do with IPV6.

Basically what you have to do is disable IPV6 AND comment it out of hosts (%SYSVOL%\system32\drivers\etc\)

This is due to a bug which you can read more about here: http://www.buit.org/2008/01/04/outlook-anywhere-is-broken-on-ipv6-in-windows-server-2008/

If you have any questions feel free to throw a comment

Good luck!

The dedicated reader will also remember the post I made when i tried to upgrade the HAS and the issues we had with the new database…. this post is related to that one.

Yesterday we received some errors in the HMS (hogia messenger service), which is basically the transport application between baltzar (which is our invoice system) and HAS. The reason was something in line of “The users password has stopped existing” or something in that matter. It was some sort of a user timeout anyway. So i figured i forgot some checkbox inside SQL 2008, and I was correct. If you create a new user you plan to keep for a while dont forget that box.. it’s located here:

Security -> Logins -> [Your user] -> Properties -> uncheck “Enforce password expiration”.

Apply and restart the services and things should get rolling again!

14:50:35 Synkroniserarversion 11.0.8161

14:50:35 Synkroniserar postlåda [NAME]

14:50:35 Synkroniseringsordning

14:50:35                1 mapp(ar) har uppdaterats i onlinearkiv

14:50:35 Synkroniserar lokala ändringar i mappen Borttaget

14:50:35 Överför till server [MAILSERVER]

14:50:35                7 objekt har tagits bort i onlinearkiv

14:50:35 Hämtar från server [MAILSERVER]

14:50:35 Färdig

14:50:35 Offlineadressbok för Microsoft Exchange

14:50:35             0X8004010F

This happened for every. single. sync.

So as you can understand this error was sort of clogging our system performance and space-wise. I tried to google, which turned up with almost nothing. Close to nothing helped me at all with this issue. After a few hours of checking (seriously) almost every hit i got on the error I resorted to do a travel inside the operating system myself. I started with event viewer, I found some issues regarding services not starting correctly. It was also complaining about my legacy exchange machine. It turns out the RPC over HTTP Proxy was the real problem however! It turned up as an error and the only ‘fix’ was to install it, so I figured what the heck, and installed it. A few minutes after I installed that service a lot of services kick-started (including outlook anywhere) and the error magically disappeared. Sometimes the windows event viewer does its job..

Down to business! When I came back i made the discovery that not all our emails were delivered the past few days. First i thought it was client-side spam filtering someone filled out by mistake but the deeper i started digging the more i realized it was something way more nasty.

Let me explain our mailflow in detail:

As you see it passes through our legacy mailbox server and then on to our exch07-machine. It took me hours and hours to search through the whole flow in detail, trailing messages and tracking where it stopped. The first break i got was when I found a message that did not get delivered to the exch07-box. It stopped at “SMTP: message submitted to categorizor”. Yes, with an o. I always thought it was spelled with categorizer.. anyway.. after a steady session of googling I found that this step is where the email is scanned in spam/virus-filters and if it gets halted in those steps it stops there.

We had an anti-spam feature installed on our legacy box. I’ve never thought about it since it didn’t change anything..  until now.

When i fixed this I started reading up on the news that I’ve missed on various tech related sites, it seems a lot of people had problems with their spam-filters after new-years. i mean seriously, y2k again? I thought we were passed this problem! Anyway, my solution was to uninstanll the trend micro anti-spam feature and now it passes through like a charm. The lost messages are however lost, sadly.

One days work lost, thanks trend micro!

So as I’ve blogged before, we moved the HAS database to a new machine because their new version required a new SQL-version. It didn’t say _anywhere_ that you had to make changes to other Hogia applications regarding it.

Anyhow, after a few days when we needed to do salary payment we noticed that we couldn’t run backups. Which is sort of weird? Two totally different applications somehow made the other one not working. Anyway it complained about rights for user ” ” “. yes one –>”<– was the user according to the application. I couldn’t figure out why it said only that one character and not the user that was logged on. Anyway I tried anything, giving the user full rights on the SQL-database, giving public full rights etc. Nothing worked. Then i read on their site that you had to download some backwards compatibility application for a new installation and I thought it wasnt nessecary because we didnt change database or move it or anything. I was like, yeah, whatever, I’ll try to set the settings manually. Finally after trying everything I figured “fuck it” and run that silly app on the client. Guess what? It worked.

Turns out that application did its job, once i ran that stuff again on the client machine it worked like a charm. The mystery of Hogia continues…

At least thats how I understand it!

We’re currently trying to configure INSIDE Backup exec to receive the data from exchange and put it on the backup box.

I think we’re done, now we can only wait for tonight when we do the first run.. If it works I’ll post a more elaborate description of what you need to do!
Peace (for now!)