However, you can get a number of ADSL connections in quickly, and users can use their corporate VPN client to reach head office.

But how to balance users across the ADSL lines? You could subnet the network and have each ADSL router as the default gateway. But that’s a lot of network change. You could also use my little technique described below.

The VBScript will read an XML file, throw a dice, and setup some static routes randomly based on the XML file. The static routes will refer to the IP addresses of your VPN endpoints.

The program then drops to shell to use a ROUTE ADD command – note it doesn’t set it permanently. Thus, the program should be set to run via a login script, or similar. Users will need to be a member of  the “Network Configuration Operators” Group.

The syntax of the XML is as shown:

<routerandom>

<rtr>

<gateway>IP.OF.FIRST.ADSL</gateway>

<route>ROUTE.OF.VPN.ENDPOINT1 MASK 255.255.255.255</route>

<route>ROUTE.OF.VPN.ENDPOINT2 MASK 255.255.255.255</route>

</rtr>

<rtr>

<route> IP.OF.2ND.ADSL MASK 255.255.255.255</route>

<route>ROUTE.OF.VPN.ENDPOINT1 MASK 255.255.255.255</route>
<route>ROUTE.OF.VPN.ENDPOINT2 MASK 255.255.255.255</route>

<gateway>IP.OF.2ND.ADSL</gateway>

</rtr>

</routerandom>

The tool is quite flexible and reliable. Unfortunately, it’s not as fault-tolerant as I would like,  because (certainly with the cisco VPN client), the software doesn’t fail over to the next-lowest-cost route if an ADSL router fails. So, if an ADSL router dies, the only option is to remove it from the XML file.

Code is here, forgive me for it being inside a word doc. RouteRandom.vbs

I ended up choosing a mac; when I specced it as I want (an i7 processor, the high res anti glare screen and faster HD) it wasn’t that much more expensive than an equivalent dell. Of course, I need to run windows on it, however, I have dabbled with the Mac side of things and have spent more time in OSX than Win7 at present. This is just a random collection of thoughts based on a few months usage.

The hardware itself is excellent – I don’t know if it’s actually more sturdy than any other laptop, but it feels better put together. I do miss the ability to have an internal 3G card, and the USB ports being too close together is annoying, but overall its easily the best hardware I’ve used. It’s quite small which is handy as I usually carry two laptops.

OSX itself I’m not especially blown away with; it’s been a few years since I used a mac in anger. I picked it up quite quickly, and the multi-touch trackpad is excellent, but the rest of it doesn’t blow me away with its super-user-friendliness or anything special. I still struggle to cope with the “menu at the top, application in a window” model of operation, and dislike the windowing maximising (or lack, thereof!)

Being able to drop into unix and find standard unix commands lurking under the surface has been quite handy. Even accounting for the fact I’m more familiar with windows, I think Windows 7 has it in the usability stakes.

As far as reliability goes, I would say they are both equal. I’ve had apps crash on both, and two kernel panics during my use.

I’m finding OSX as a good base to run Virtualbox, I have no evidence but VMs do feel quicker running in OSX rather than windows.

In terms of apps, I’ve found everything I need, apart from Office and Visio. I haven’t invested in MS Office for mac yet, and I think open office & Neo office are lousy handling word documents (I’ve timed it at over a minute to open a word document – unacceptable. I’ll look carefully at office for mac before making my choice but for now continue to get into Windows and Office/ Visio 2010

I’m dealing with the odd keyboard layout easily, my brain seems quite comfortable leaping between the mac keyboard and a normal keyboard.

Overall, what has surprised me is how little the difference is between the two. Any dev work I do is in VMs, and other than the office issues discussed above, I struggle to have a compelling reason to use one over the other. Im sticking with OSX just now, more for the novelty factor but think long term I’ll find myself in Win7 more.

However, times are different now, there is no money left. There’s a lot of security controls in place to mitigate against risks. Is it time to accept some of these risks and pare down the controls ? Let’s look at what can be done, comparing against a normal large business as our “sanity control”. Similarly, most businesses do not have to deal with “life and death” information, so I’m not considering that classification of information.

Government policy will not allow a wide area network to be run without complex encryption over the top. So, whilst most companies buy MPLS from the likes of Cable and Wireless and BT, government will do the same, then overlay a complex and expensive VPN. Removing this as a mandatory requirement would reduce costs in future, and even for currently deployed networks there’s a need to support all these extra boxes (and give them power, cooling). Plus, removing the extra encryption would improve speeds! As long as the migration cost is less than the support cost, everyone wins!

There’s a requirement for hard-disk encryption. Most corporates have woken up to this as an issue, and central government is no different. However, rather than effectively mandating a few, expensively approved products, perhaps the use of common commercial alternatives would save tens of pounds per machine.

VPN is another common business operation. Again, common in Government but mainly done with exotic VPN products you’ve never heard of. Ditch this, and go with Juniper and Cisco that everyone else uses. Many corporates will provide webmail for their employees. This will allow employees to access their email, probably from their home PC. This might alleviate the need for a blackberry, laptop and so on. You just won’t see this on a central government system. So, provide this and see mobile comms costs tumble.

Each government department is an autonomous organisation. They are joined up via the “Government Secure intranet”. This is a private WAN used to ship email and allow access to each others private websites. For email, if you’re feeling bold you could enforce TLS between your partners, or have a select few use PGP. But use the internet like everyone else. And when business want to share information, they setup VPNs over the internet. Do all this, and scrap the GSi

You’ll note none of these suggestions are fundamental. I’m not suggesting everyone run linux, or some sort of single, unified IT system. Mainly because change == cost, and drastic change == lots of cost.

However, there has to be a downside, and that is risk. Our attackers will have an easier ride, and those who seek to get at our information will have more success. As cyber-terrorism becomes a reality, would we be setting ourselves us for attack? At the more trivial end there’s bound to be stories about Nigerian scammers getting into government accounts.

The controls that are in place are not there because some security nerd wanted to install the latest gizmo. The question is therefore, is there anyone senior enough to take these decisions and also genuinely accept the risks and guaranteed issues?

Below are a series of suggested tasks to get you up-and-running in the IT infrastructure world. Intentionally I’ve not explained every step in great detail, nor included everything you have to do. Nor will performing these tasks make you an expert in these technologies. In fact, one of the goals of the exercise is to get to comfortable and familiar with new technology, googling for information, doing some “try and see” in a safe environment

1. Build yourself a PC
In years gone by, building a PC from components was a good way to get a cheap PC. These days, less so. However, we have particular needs from this PC, and the actual building and fault-finding process will help us along the path. Exact spec is up-to-you, but we need:
• As much RAM as possible (suggest 8GB)
• Processor capable of 64 bit OS and virtualisation
• DVD Drive
Otherwise, it needn’t be the highest spec. You should check all drivers are available in 64 bit versions, however happily it’s very unusual these days for this not to be the case

2. Microsoft Technet Direct
A Microsoft Techet Direct subscription is something every windows techie should have. For just over a hundred pounds a year, it gives you full access to all Microsoft business software, and is great for testing and evaluating –just as we’re doing here. So get yourself a subscription and make the first thing you download Windows 2008R2 as we’re going to build a ..

3. Virtual environment
Now we’ve got a shiny new PC, lets start to do something with it. Burn your Windows 2008 R2 to a DVD and pop it in your machine. Build the OS as you see fit, and have the Hyper-V role installed. We’re going to use that as our virtualisation software. Other than basic software you need to manage the server (I’d also suggest 7-zip is a good tool), you shouldn’t modify the base server. That’s what VMs are for!
First things first, let’s build a basic server image VM. Fire up the Hyper V console and configure it up with settings you think make sense. Copy the Windows 2008 R2 ISO file to the machine and mount that. Turn on your virtual machine and install Windows 2008 R2. When it was finished building, ensure you install the Hyper-V tools.
Close the virtual machine down, and take a copy of the VHD. We’ll use that as a “gold” image to build other hyper-V machines.

4. Build an Active Directory
Our first server is going to be an Active Directory server – this is used by almost all other windows server system components so makes sense to build first. Copy the Gold VM VHD and configure a new VM – I’d give it say 4GB of RAM whilst we’re using the machine and when it’s just running in a steady state reduce the amount of RAM.
Use the NEWSID tool to ensure the machine has a unique identifier.
Installing this will also install DNS – setup forwarding so that it forwards to your ISPs DNS servers.
Decide on the structure of your OUs, where you will put users, computers and servers
Create some users and groups called
• Admins
• Managers
• Staff

5. Install WSUS

This might be a learning environment, but we want to follow best practise! So download WSUS from Microsoft. I’ll leave it for you to decide if you want to install it on the Active Directory Sever, or build a new server to host it.
The next thing to do is to build a GPO to ensure all machines refer to the local WSUS server for updates. Decide on your update strategy both in terms of WSUS approvals and application of patches. I’d be included to have automated approvals and install as much as possible as this is only a trial environment.

6. PC Image
If possible, this should be on a “real PC”. If we don’t have the kit, then a virtual machine will have to do. I’ll leave the operating system choice up to you, but XP is still a valid choice as it’s still used everywhere – although it might have added complexity with your automated deployment tool.
What we’re doing here is building a PC in the anticipation that it’s going to be rolled out to thousands of desktops. So we want the initial install scripted (i.e. automated input of user details, serial number and so on).
Include any drivers that your target machines are likely to need, service packs and patches. Don’t install any software (that will follow)
Then, follow the instructions for preparing the machine for imaging, which will include resetting the SID, wiping the swap files and so on.
You need to decide on a deployment method: RIS or WDS. WDS is the newer technology but there might be a reason you want to choose RIS especially if you have XP as your OS.
Once you have that up-and-running, image a few PCs (virtual or real) and see how you get on.

7. Install Software
Most big companies will have a heavyweight deployment tool to package and deploy software, here we’re going to keep It simple and use the builtin windows software deployment.
Download some Microsoft software (suggest Office and Visio) and configure these packages so AD will deploy it to all PCs (not servers though!)

8. File and Print Server
We want to setup a file share with particular access rights..
This should be
• Admins – Full Control
• Managers – Change
• Staff – Read only
Also, all users should have this drive mapped as their “X: Drive” upon login automatically.
It’s your choice whether to setup another dedicated file server VM or “piggy back” upon another one.
Your next task is to setup a network printer. This should be
configured so that users can connect to \\servername\printername and have drivers for that printer automatically installed. Note if you have a USB printer it may well be easier to share this from the "real" server

9. Exchange
This is a big one! I would actually suggest installing Exchange 2003 as many companies still use it, and migrating away from it is a useful exercise in itself. However, your gold VM image will not be sufficient as Exchange 2003 needs a 32 bit OS.
Build a new VM, Install Exchange 2003 and create exchange mailboxes for your users.
Now here comes the clever bit. We’re going to setup email routing to and from the internet. Go to a provider of dynamic DNS services like dyndns.com and setup a DNS name for your organisation that’s registered against your current connections IP address. Now, also setup an MX record to the same address. You now need to configure your ADSL router/ Cable modem/ etc/ to forward port 25 traffic from the internet to the IP address of your Exchange Server
Automatically create email addresses for your users in the format of name@your-dynamic-dns-entry
Finally you should configure outlook so that it automatically creaes a profile for end users to connect to the their new mailbox.

10. Document
Now that we’ve got a cracking IT infrastructure, let’s have a go at documenting it (OK, we should probably do that first, but hey, this is only an exercise. Fire up visio (downloaded from your technet subscription) and describe your environment. Your diagram should include
• All your servers, names, IP address, function
• Active Directory
• Exchange
• Internet connection
• How mail is routed in an out
• Virtual versus real machines

The need was for a “point in time” inventory of what’s out there. A we didn’t have use of a big tool like centinel nor wish to use the existing network management tools (to provide an independent result). Oh, and I had limited admin access rights.

Here’s how I did it

Largely, the work was split into two pieces, a Wintel Audit and an “non Wintel” – encompassing networks, printers, switches, SANs…

The Wintel Audit was fairly easy – I cobbled together a VBScript to query machines using WMI and pull back installed software, machine spec, and so on – just the basic in you might want if you need to take migration decisions. I’ll post the script up in my next blog entry.

The Non-Wintel was more involved. Firstly, i used nmap to scan every IP device. It takes an “educated guess” as to what the device is, and does a reasonable job. The most surprising fact was there was quite a lot wintel kit in here I hadn’t picked up. This was because machines were in different domains and workgroups. These were then added to the wintel audit.

This gave me an outline of what to look for and how to investigate.

There was hundreds of printers on the estate, almost all HP. The nmap tool had done a reasonable job of guessing the type, but it wasn’t precise. To do this,  I fired up HP Jet Direct tools, which is a little light-weight tool that HP no longer provide in this basic form. Shame, because it’s all that’s needed. I just gave the IP addresses relating to HP printers and it went off an queried them. Minutes later, I had netbios names and proper printer models. Lovely.

I didn’t have full access to networking devices, but I did have the SNMP community strings. Therefore I used Billy the Kid’s cisco SNMP tool.

I simply fired in the switches IP address, the community string and the tool got the switches CDP neighbours, helpfully giving me the model names and IP addresses of Layer-2 connected switches. From this, I was not only able to build a network map,  I was able to make the inventory far more accurate.

However, there was an area that was, by design, hidden from view. The client has multiple connections to multiple systems, so has a myriad of firewalls and DMZ’s. I peered through the firewall rulesets to see if I could find equipment on the network that was hidden from ICMP scans. Easy on the Cisco ASDM, Checkpoint FW1 and the Juniper – slightly more complex reading the config of an old PIX! Doing this enabled me to find servers, switches and more Firewalls behind Firewalls.

Then is was just a case of manually picking off the oddities. The nmap scan found lots of Sun boxes, helpfully for me they all revealed their machine name when I FTPd to them,  or finger @machinename. Almost all other devices tell me enough to be useful by connecting via Telnet, SSH, http or https – APC power strips and specialised printers. I even found an plasma screen that talks IP!

The result? An asset list that’s about double the previous list…. and a lot of “must improve housekeeping” to do !

It would try to login, try for a while..

Before dying with Error Code: 8000402a:

Clicking “Get more information” was final insult as all the results were in French. Sacre-Bleu!

3 supply a private IP address and there’s been a number of occasions where the web has been dead, but TCP/IP connectivity is fine, therefore I assume they use a “transparent proxy”.  Which I suspected of being at fault. Not much I could do about that, and didn’t fancy calling the call centre whilst I was on the train

Googling was difficult  – When you are a company selling 3G services, calling it “3” isn’t helpful!

The diagnostics built into messenger wasn’t very helpful, apparently I have a Invalid IP address. Works OK for the web.

And, in fairness, the IP address arrangement handed down by the 3G connection do seem a bit odd..

But this lead to to a dead end.. so what now ?

• Firewall… Tried that, no difference
• Windows Event log – nothing
• Diagnostics in Live Messenger – too cryptic for me to decypher

My next steps were to fire up wireshark and try and understand what’s going on. But, sometimes inspiration comes from funny places

I decided to see what would happen if I run Live messenger in Admin mode. No difference.

Then I tried compatibility mode, changing it to Windows XP SP3:

A UAC pops up, I select “YES”

And…..drumroll….. SUCESS!!

All I need to do now is configure UAC not to complain when messenger Launches!

However, it appeared to be buggy and didn’t delete all files. No obvious reason why. Also, if the temp area wasn’t available when it was run, it would error to the screen.

To help, I concocted this script. I’m no VBScript guru, but with looking at examples on the web pulled this together. It majors on letting the user know what it’s doing and trapping errors. The program style isn’t perfect, but it works.

Most importantly of all, we found the bugs in the previous effort; some outlook attachments were marked as read only and therefore not deleted; this script gets round that by forcing deletion – only stopped by permissions problems.

Here it is in all its glory..

Option Explicit

‘ Path to delete
const FolderName ="d:\Temp"
‘ Age of file (last modified) in days before it is deleted
const Days=7

‘Set debug to 1 for screen pops with helpful messages
const debugmode=0

const SUCCESS=0
const ERROR=1
const WARNING=2
const INFORMATION=4
dim result
logit "Begining process to delete temp files in " & FolderName & " older than " & Days & " Days",INFORMATION
result = delFiles(FolderName,Days)

Function delFiles(strDir,strDays)
‘ Take in the path to delete and the number of days old a file has to be
‘ Then delete the files in that path if they are older than that date

  dim fso,file,folder,individualFile,foldercontains
  dim strComments
  dim intDeleted, intNotDeleted
  intDeleted=0
  intNotDeleted=0
  Set fso = CreateObject("Scripting.FileSystemObject")
  on error resume next

  Set folder = fso.GetFolder(strDir)
  if err = 0 then
    strComments = strComments & strDir & " exists." & VBCRLF
  else
    strComments = strComments & ". **ERROR** Accessing folder – cannot find " & strDir & VBCRLF &  err.description & " (" & err.source & ")" & VBCRLF

    intNotDeleted=intNotDeleted+1
  end if
  err.Clear 

  Set foldercontains = folder.Files
  dim intDifferenceinDays
  strComments = strComments & "Deleting Files older than " & strDays & " days" & VBCRLF
  For Each individualFile in folderContains
   ‘ Loop through each file in the folder and check its date
    if debugmode=1 then
                  wscript.echo ("Looking at " & individualfile & VBCRLF & "Which has date last modified of: " & individualFile.datelastmodified _

                                  & VBCRLF & "To see if its " & strDays & " days older than " & Now)
                End if
    strComments = strComments & VBCRLF & "Analysed " & individualfile & ": "
    intDifferenceinDays = DateDiff("d", individualFile.datelastmodified, Now)
    If intDifferenceinDays > strDays Then
                  if debugmode=1 then
                    wscript.echo ("We’ve decided to delete" & file &"… Datediff is" & intDifferenceinDays )
                  End if  
      strComments = strComments & " Deleting…."
      on error resume next
      fso.DeleteFile individualFile,TRUE
      if err = 0 then
                    intDeleted=intDeleted+1
                    strComments = strComments & ". SUCCESS"
                  else
                    intNotDeleted=intNotDeleted+1
                    strComments = strComments & ". **ERROR** " & err.description & " (" & err.source & ")"
                  end if
      err.Clear 
     Else
       if debugmode=1 then
                     wscript.echo ("We’ve decided to spare" & file &"… DiD: " & intDifferenceinDays & "  Required: " & strDays)

                   End if
                   strComments = strComments & "Not Deleted. Only " & intDifferenceinDays & " days old"
    End If
  Next
  strComments = strComments & VBCRLF & "No of Files Deleted: " & intDeleted & VBCRLF & "ERROR in deleting: " & intNotDeleted

  if intNotDeleted > 0 then
    logit strComments, ERROR
  else
    logit strComments,INFORMATION
  end if
  delFiles=1
End Function

Function logit(text,level)
‘ Writes a simple message to the windows event log
  dim Wshshell
  set WshShell = CreateObject("WScript.Shell")
  WshShell.LogEvent level, text
  logit=1

end Function

For any previous Exchange deployment (certainly 2003) you’d start with a SAN and use features like snapping to ensure you can backup without affecting performance.

To my mind SANs remain stubbornly expensive for what’s actually delivered (I was just quoted over £3500 for a single 15K 600GB SAS disk which is certified to run in a SAN!).

So the fact I really don’t need one for Exchange 2010 is perfect.

But how do I back up?

Microsoft will tell you they don’t back up at all, just rely on the native protection and deleted items retention.

I’m a little – just a little- less gung-ho than that and I suspect many of my customers are, too.

There’s very little product choice on the market, or indeed much Microsoft collateral about how to backup Exchange 2010, so I thought I’d take a stab at a possible solution myself!

My objectives:

• Don’t “waste” the genuine data protection tools native to Exchange 2010
• Prepare for the day when an exchange corruption is replicated to ALL databases (however impossible that might be)
• Provide a longer-term archive.

Consider the following environment:

We’ve got a single DAG with four copies of every database. Let’s say for argument sake we’ve got 5000 users with a 1GB mailbox. Of course, our disks are directly attached and we’re using nice-and-cheap SATA storage on JBOD. Let’s use 1TB because smaller disks are beer money less.

So far, so good, so like every other Exchange 2010 design piece. We’re leveraging the native protection and we’ve got four copies of the data.

But how to protect against the replicated corruption scenario?

I’m using another new feature of Exchange 2010; lagged replication. So this server in question is always behind the other servers; in theory then should the “replicated corruption” scenario occur, we can take action before it plays into our time delayed mailbox server.

But how long? Too short a delay and the corruption might get missed and played into the lagged database anyway Too long and and invocation of the lagged server might risk losing mail.

My best-guess figure was about 24 hours; this is comparable to a normal restore if we don’t have logfiles.

Now, observant types will have noticed there’s extra disk arrays attached to the lagged mailbox server. To break with custom, these will be RAID5 and their purpose is to act as a file share area to perform backup-to-disk operations. I’m doing disk-to-disk backups because:

I can, at very little infrastructure cost

Having recent backups online is always useful.

At the time of writing, the choice of backup products is underwhelming so I’m going to use the built-in tool. The real downside to this is that I can only backup from the active node, thus I need to be real careful about what I’m backing up, when. Pumping the data across the network in good time might be tricky without the right network setup.

Most likely, one or two databases will get backed up every night with all databases having at least an incremental backup

Now to the final part of the plan; the long-term archive. Hopefully never needed, but your operation might need to keep archives of data (this, probably isn’t the solution for this, you need to check out other new exchange features). But it’s most likely needed when the CEO needs an email he deleted 12 months ago.

Backup-to-tape is therefore meets my need. I’m only going to backup to tape the files produced by the disk-to-disk backup process, and I’m going to choose my timings wisely.

So there-we-have-it. A fairly robust backup architecture? I’m hoping as time progresses and products fill the void (like DPM2010) this solution will look archaic, but for now it’s my best shot at what backup could look like.

Have you ever been in a scenario where you need to change a password on a service account but don’t know what service on what servers use the account? You could pick through audit logs and it still might not tell you if a service hasn’t been restarted recently. Regscan will visit all machines in your domain and give you a list of machines that use that account

Usage

Simply enter

regscan account domain [textfile.txt]

where:

• account is the account you are searching for. Don’t put the domain name first, regscan will pick out either notation from the service list
• domain is the netbios domain name to search
• textfile.txt (optional, but reccommended) Specifies a list of servers to search, one per line. In large domains, this is a more reliable method than leaving the program to scan the domain to find machines.

Download

Grab the program here. Let us know how you get on with it.

Spotted on a couple of my machines, windows update was not working, with the above error:

The Microsoft TechNet article is pretty unhelpful, suggesting the windows update service is having trouble connecting, possibly an on-machine firewall stopping it.

Nothing that should be stopping this springs to mind, so my first concern is malware.  A quick scan by Malwarebytes didn’t show anything; sadly I know that doesn’t guarantee we’re OK. I had a quick look at the host file; nothing changed there. The IP addresses associated with the windowsupdate DNS names appeared to be OK. It did seem as if the PC was being blocked from geting updates.

So, what is actually happening when I click “Get updates” ?

I needed something to let me see behind the lovely chromed update UI. The tool I chose was was fiddler. Mainly used by people debugging websites, it also has the useful knack of sniffing all http traffic from the machine. Let’s fire it up and hit the “try again” button:

We can see the update process requesting the wuident.cab from a server jelly.dessert.local

clearly, the machine in question doesn’t belong to WindowsUpdate. Fortunately, there’s an explanation which is less worrying than some uber-weird virus.

A few weeks ago, I need a couple hundred GBs of disk space for some new VMs in a hurry. Being in a tight spot, I uninstalled WUS which conveniently was taking up about that much space; I then of course changed group policy so that my dozen or so  machines talked to windows update directly

It would appear, however, that a couple of machines have group-policy update issues and never got the update changing from using a local WUS to the microsoft update servers.

So a fairly predictable fix from there on in. But the original fault-finding would be soooo much easier with a little more diagnostic error messages, Microsoft!