It is now easy to build, and customise all parts of your OS deployment using these technology’s, but as I started using it there was one thing missing from it that old style RIS used to have.  Incremental naming of machines as you build them.

In RIS is was a simple matter of putting a variable into it’s settings, and it worked no problem, now however even though there are hundreds of new options naming in this manor is not one of them. I did a search around and come up with this article at Michael Niehause’s blog.

I have re-written the code from this posting in vb.net, i also had to make a couple of changes to my rules file to get it to work. The only differences between my code and Michael’s is mine is in vb.net and his is c# and mine does not add the machine account with a description to the active directory, I just let the install process do this for me.

The Code

Fire up a new web service project, and add this function to it:

<WebMethod(Description:="Returns a new Computer Name to use." & _
" Inputs dns domain, a prefix and OU")> _
Public Function GetNewName(ByVal DnsDomain As String, _
ByVal NamePrefix As String) As String

'build the search, based on the passed domain, then the name prefix
Dim adRoot As New DirectoryEntry("LDAP://" & DnsDomain)
Dim dirSearch As New DirectorySearcher(adRoot, "(name=" & NamePrefix & "*)")
Dim existingNames As New Dictionary(Of String, Guid)()

'loop through the results
For Each result As SearchResult In dirSearch.FindAll()

'get the name, and create a guid holder
Dim strCurrName As String = result.Properties("name")(0).ToString.ToUpper
Dim netbootGuid As New Guid

'if we have a guid use that with this name
If result.Properties("netbootGuid").Count > 0 Then
netbootGuid = New Guid(DirectCast(result.Properties("netbootGuid")(0), Byte()))
End If

'add details to the dictonary, if they are not already there
If Not existingNames.ContainsKey(strCurrName) Then
existingNames.Add(strCurrName, netbootGuid)
End If

Next

'now try and get the next name in sequence
Dim strNextName As String = Nothing

'loop through all the avalabe machine numbers up to max of 999
For i As Int32 = 1 To 999

Dim strNameTest As String = NamePrefix & i.ToString("000")

'is this name in the list, if not we have our name
If Not existingNames.ContainsKey(strNameTest) Then
strNextName = strNameTest
Exit For
End If
Next

'return our new name
Return strNextName

End Function

What its doing

All the code is doing is querying your active directory for machines named like you specify, that end with a unique 3 char number (from 001 to 999). Once it finds one it returns the new name back to you. The good thing about this code is it will re-use any machine you have retired, so no numbers go to waste.

You pass the function the dns name of your active directory and the first part of the computer name (e.g. CA-HR-), and that’s it.

Calling the naming function from MDT

Once you have done your web service and it is callable from your web server, you need to alter your MDT rules file to call it and get a new name. This is all the sections you will have to fill in to get it to work, you will probaly have a lot more in your production system:

[Settings]
Priority=Default,GetNewName
Properties=DnsDomain, NamePrefix

[Default]
SkipComputerName=Yes
DnsDomain=mysite.comp.co.uk
NamePrefix=CA-HR-

[GetNewName]
WebService=http://your-web-server/WDSExtras/WDSNameExtras.asmx/GetNewName
Parameters=DnsDomain, NamePrefix
OSDComputerName=string

All that is happening here is you are telling MDT to run an extra section called GetNewName, and that section is going to call your web service with 2 parameters (created in the default section) and then set OSDComputerName (the key that holds the machine name) to the returning string.

Troubleshooting

Odds are you wont get this right first time, but MDT writes does quite a good job of logging what it has done. On a machine your have deployed if you look in C:\MININT\SMSOSD\OSDLOGS you should be able to find the section that is trying to call the web service, and see what it did, and any errors if they occurred

Its a simple plugin that will allow you to use a directory of images you have on your web space as a simple gallery on page in your blog. There seemed to be a lot of plugins out there for getting images from flickr and other image hosting services, but not much if you wanted to get at the images your self.

Any comments, suggestions etc please let me know!

I haven’t used Spiceworks since version 2, I was quite impressed with it back then and I know the product has developed quite a lot in that time, so I was hoping it has improved even more. As in the previous post i have done a few timings, and noted what the scanning of the network has managed to find.

This test was completed in a completely different environment than the previous tests, so its not really fair to compare this version with say version 2. Spiceworks was installed to a Dell 1850 server, running windows 2003 with 2gig ram. I scanned 6 subnets in all, containing lots of different devices (pc’s, phones, routers etc). Clients are mainly Dell optiplex’s running Win XP, printers mainly HP.

Test Results

Test Conclusions

Speed
The first thing to hit me was the scanning speed, I had given spiceworks a lot more work do to then in the previous tests, but i was still a little disapointed at the time it took to initially scan the network. There was a lot of kit to find and identify, but it took a little longer than I would have thought

That said, the initial scan is a one of event, and subsequent ones once the majority of devices have been found are nice and fast. I’m not sure if the speed is really that big an issue, I would prefer to sacrifice a bit of scan time to get more accurate results.

Hardware Results
So, did Spiceworks actually manage to identify all the hardware on the network, and work out what it was. For the most part, yes it did. Unless there was a connectivity problem with the machine, it did manage to work out what it was. I correctly identified all the workstations and servers, and managed to extract out the information on the hardware.

It does tend to class most of the network kit as ‘Other Devices’, i’m assuming this is because it cant authorise its self on the device, but i’m not sure, it just seems to see them as http devices. I was going to try and scan a test device, but there is no scanning option to pass a telnet password to a box, even if you could I’m not convinced it would be able to tell what it was

It did however, manage to recognise all the printer correctly, and in most cases it was reporting back information on the status of the device correctly

Software Result
This is one area where I found it difficult to tell if Spiceworks was working right or not, as I dont have an up-to-date list of the software on the machines. To try and test it I picked a couple of machines at random and looked at the software on them, it managed in both cases to get pretty close to what was on the actual box. I like the feature of the software list that tells you the oldest and newest versions of things, this would be very handy if you are rolling out some software and want to make sure everything has got the correct version

Interface and Features
As I have not used Spiceworks since version 2, i’m not sure what is new this version and what has been in for a while. I do like the time line function, it would be nice if it could update in real time but then you would be scanning your network all the time.

The reporting system seems a lot more customizable, as you can now not only decided what you want to see, but organise the columns on the reports as well, and filter down what you want to see. The export options are pritty standard, but I would have expected to see an xml option along with the others. It is missing some advanced reporting features (grouping, custom fields etc.) and has no support for generating graphs, but those are probably not that important.

The network map is a nice idea, but i’m not really sure what use it would be, normally you would probably only be interested in this information for your network kit, and Cisco already have some specific tools to do this.

Summing Up
While the product is not perfect by any means, for a free bit of software you cant really get much better for a small or medium size organisation. The hardware/software inventory is accurate, if a little slow and does seem to struggle with network kit.

The interface is very easy to use, and its easy to customise the dashboard so you can see the information that interests you. The adverts do get in the way a bit sometimes, but you can always pay to get them removed.

All in all this is another very good version, and if your in the market for some network auditing/reporting tools, for the (lack of) price this is one of the best.

SQL reporting services has been part of the SQL Server suite for a few years now, and this book covers the latest version 2008. As stated by the back cover of the book, it is aimed at beginners, but also has plenty of useful information for old hands, with all of the topics covered in the book backed up by code examples and screen shots.

As with most books I have read on any reporting engines, a basic knowledge of SQL server, and the SQL language will always help a great deal when reading the book. Also a couple of the topics hit on using the reporting service from your own applications, so you will need some programing know how to get anything from these chapters.

The book is written in a very easy to read way, and unlike some technical books manages to seems light enough to enjoy reading, but everything is still explained in an accurate and concise manner.

There are 2 sections given over to using the new report Builder (something i have never used before), explaining how it is set out like excising Microsoft office application, and information and examples of all the other new features in this version. The section on using Crystal reports is interesting, and although i found it a little short, it is something i have never seen before in a SQL reporting services book, they never seem to touch on the subject at all.

The chapters at the end of of the book are about programing into/around the reporting engine, I found the one about Progammatically Creating a report particularly interesting, it takes you through all the necessary steps to create your own .rdl, using the XMLTextWritter class. Might not be very interesting if you just want to author reports in the builder, but if you wanted to take it to this level the information is very helpful, with full source code in the book.

As well as all the new features for 2008, all the existing one are covered (images, sub-reports, parametrised reports, datasets, formatting, graphs etc) in good detail, with examples and screen shots available for each topic.

All in all not only enjoyed reading this book, i took quite a bit of useful information out it as well, some of witch i think i would struggle to find in other publications. A good read all round.

Buy this book from Packt.com now!

Parameters are you friend

The first thing to do is create yourself a new parameter in the report, and call it something like par_expand_groups. Give it the following options:

Now, you need to make this parameter control weather or not the group is expanded as the report is ran, you do this by putting an expression into the hidden property of the row.

Normally if you have a group, when you run the report it will probably be hidden and the user has the option of showing the details, so the hidden property will be set to true. We are going to set it to the following expression:

=IIF(Parameters!par_expand_groups.Value = “True”, False, True)

This means that if the parameter is set to yes, the details will be shown, if its set to no they will remain hidden. Add this to each level of grouping on your report, and the user now has control over what they see

Slight Problem with the show/hide icons

Once you have added this to your report, you might notice a slight problem. If you run the report with the option of showing the group details, the little show/hide icon on the left used to toggle the details will be wrong. It will be showing the ‘+’ even though the details are already showing, and if you press it to hide the details it will change to a ‘-’. They are the wrong way round!

Obviously, this is going to confuse your users, and needs sorting. Well, thanks to a tip from http://www.lukehayler.com/2009/07/custom-visibility-toggling-in-ssrs its easy.

All you need to do is open up the properties of the text box that is the trigger for the show hide, and go into the Visibility tab. At the bottom is an  option to set an expression for the toggle image. just enter:

=IIF(Parameters!par_expand_groups.Value = “True”, True, False)

As before, they is just looking at your parameter and setting the value of the toggle image. Now, when you run the report with the option to display the details, they icons for toggling the details should be correct.

I while ago I wrote an artical (here)about the new Web slice feature in IE8. A few people have asked about a plugin for WordPress, and the other night I finally decided to have a go.

Head over to http://www.beakersoft.co.uk/wordslice for all the info, or to http://wordpress.org/extend/plugins/wordslice/ for the download and the install notes.

Hopefully in the not to distant future i will be add some extra features to it (its kind of limited at the moment) so stay tuned!

I have recently had the unfortunate experience of having to try and clean up a network that has been hit by the conficker virus.It is a pain to get of and caused us a lot of problems, this is a quick guide to how we went about cleaning it up. This might not work exactly the same for you but it might give you some pointers.

Accounts locking out

This was the probably the biggest pain for us. One of the way the virus spreads it’s self is on smb shares, to it tries to access these shares by using accounts it finds on the active directory/local machine. It tries to brute force password crack the accounts, and if you have an account lockout policy in place (i would imagine most places do) then the accounts it finds will become locked very quickly.

To get round this rather large issue temporally we got a script that we ran every minute or so on each of the active directory boxes, it basically just goes through the entire domain and unlocks all the accounts. We ran it in the task scheduler as two separate users in case one was locked when the task ran.

Obviously this isn’t a long term solution and should be removed once you are reasonably happy its no longer locking accounts.

You can download the script here

Finding The Infections

There are quite a few tools kicking around to find the infections, that all work with various degrees of success, the main tool we have used is one from eEye digital security, you can get it from http://www.eeye.com/html/downloads/other/ConfickerScanner.html. This one will scan a subnet and tell you what is infected and what is vulnerable (un-patched).

You can also use a beta version of nmap, this one also tells you what it thinks is infected and what is vulnerable to the virus, info about it at http://www.skullsecurity.org/blog/?p=209

We have also used an in house tool that logs all the account lock outs into a database to track down the actual machines that are locking users out, this has been the most pro-active way of finding infected machines. You can read more about the system we use in a previous post here

On the actual infected machines there are various pointers, quite a few of the services get stopped (these include server, workstation,bits,windows update), the machine can end up being completely unresponsive due to the virus hammering the processor/memory.

Also on machines that aren’t necessarily infected you might get lots of new scheduled tasks Called AT#1, other infected machines on network will have probably created these. Most of the clean up tools don’t seem to delete them you have to kill them manually.

Cleaning the Infections

Once you have identified the machines infected you will need to clean them. In theory the anti virus software should just be finding it and putting the virus into quarantine, but in a lot of cases that’s not the case. When we found an infected machine this is the process we have used.

• First kill of the memory from memory. We have been using the Symantec  W32.Downadup Removal tool available here. Re-name the app to something random so the virus doesn’t just kill it of, then run it. Once its got passed processing the apps in memory, cancel it. It will complain its not checked all the files, but it should kill the process out of memory and remove entries to it in the registry so it wont start on next reboot. It will ask you to reboot to clean up everything else but dont yet.
• Before rebooting make sure you have got the patch on, download it from here. Note that it appears you need to install at least service pack 2 on xp, 4 on 2000 and 1 on 2003 for the patch to install. Once you are patched reboot the machine
• When the machine comes back, hopefully the virus wont have loaded back into memory. Update your virus software with the latest sig files (or install some if god forbid there is none installed!), then run a  full scan of your machine. This should find the files, and Quarantine them, allowing you to  delete them.
• Look in the scheduled tasks and remove any that are called At#1 etc.

Hopefully now, even if the machine gets the virus again the anti virus software will pick it up and stop it doing any harm.

If anyone else has anymore info for detecting/cleaning this beast up feel free to email me or leave a comment and i will update this post

While debugging the app, i came across a problem that i thought i was catching and trapping, but my app seamed to be ignoreing the problem and carrying on running so i was getting some strange results.

I was running some sql similar to this:

SELECT Count(title) as rec_count FROM tbl_text WHERE tbl_text.Title=’WXXX’
AND Contains(tbl_text.page_text, ‘Tax Evasion’ AND ‘Joe Blogs’)

In the sql i had constructed i just had an apostrophe out of place (i know rookie mistake), so the sql was failing to run. Problem was the try/catch loop it was in did’t seem to mind! I got out the sql and ran it in sql advantage, and sure enough i got this error:

Server: Msg 7631, Level 15, State 1, Line 1
Syntax error occurred near ‘Evasion‘. Expected ””’ in search condition ‘Tax Evasion’ AND ‘Joe Blogs’.

So sql advantage managed to trap the error, but the dotnet framework for some reason took no notice of it and carried on running. I got round the problem being checking my record set had rows before trying to use it, but i have no idea why an execption was not rasied in the first place. Anyone any ideas?

Me and a couple of the guys i work with a decided to have a good at writing our own software, and making it available online. We are going under the (rather good i think) name of we3soft, and you can get at our website by visiting www.we3soft.com

We have got quite a few ideas/applications in the pipeline, and hope to start working with some of our local businesses on some of there projects. At the mo we have only one application (and its in beta) avaialble for download on the website. It goes by the name of SwitchWatch.

Basiclly, you can use it to tell you what ports in cisco switches have been unused for a period of time you specifically, this will allow you to tidy up the rats next of cables that lives in your comms room. At the moment the application is still in beta, and will only report on the first 5 unused ports per switch. In order to encourage people to test the app, we are giving away 20 free licences once it comes out of beta.

All you have to do to quailfy for a free copy is head over to www.we3soft.com/switchwatch, download the app and tell us what you think about it in our forum. Let us know if it worked fine, if you had any problems with it or if you have any feature requests. If your are in the first 20 to give us your comment we will send you a free licnece key as soon as we come out of beta, so head over there now!!

‘Cannot load client print control’

Turns out that one of the patches Microsoft issue out (on a seemingly never ending basis) breaks(well, blocks) the activex. The patch in question is the ActiveX killbit one (i cant remember the kb number), there appears to be a problem with the old version of the print control activex, and when the latest patch is applied to the clients it stops the print control working.

The way to get it working again is basilcly to patch your reports server. This is something i am always a bit cautious of doing, as in the past i have had problems (service pack one removed the ability to do a multi select parameter) but this time i had a virtual server i could snapshot, patch then test.

To get it working again, i installed first of all SQL 2005 service pack 2 (http://www.microsoft.com/downloads/details.aspx?FamilyId=d07219b2-1e23-49c8-8f0c-63fa18f26d3a&displaylang=en), then the latest hotfix rollup (http://support.microsoft.com/kb/956854). On a couple of the client machines this seemed to fix the problem without having to re-download the control, but on others we had to remove the control first, then re-install it and it worked a treat