Cool Windows 10 update.. and there goes my network … NOW WHAT? #lifehack

windowslogoFor those also a edgy as me, and thus living on the Windows Insiders Fast Ring might come to a surprise from time to time. Today i was passed on the latest build (14279), and of course very happy with it Smile (geek as I am). Unfortunately that wasn’t counted on networkadapters, Hyper-V and the update itself.


So what happened: I had a full setup of Hyper-V active on the machine, with a couple of vAdapters and a vSwitch belonging to that. After the update and the reboot, i all of a sudden could no longer connect to the internet. What happened? I don’t get it… i checked the Wi-Fi connection … seemed connected.. although now i was getting a 169.x.x.x loopback address (you know,  the ones that come for free Sad smile )


I went to check out the adapters … and guess what , my network bridge was gone …. weird … So i decided to remove it from the adapters .. but … COMPUTER SAYS NO ..

All of a sudden my adapters start misbehaving and havig an attitude against me.  Luckily i remembered someting from the old day … netcfg …. Smile



so what do you need to do:

  1. open a command prompt in Admin Mode
  2. run netcfg /? to get you all the netfcg help info. The key lies in the netcfg –d, as this will reset and remove all adapters and put them back in zero configuration
  3. run the command with netcfg –d
  4. reboot the device

The only thing now of course is to re-create the vSwitch and all should be back to normal.


It’s not the best solution around (it shouldn’t be happeneing at first), but still it does the trick.

I hope i could save you all some time here



The nightmare before Christmas … RESOLVED #ALM #TFS

logo_vs-v_web120I have to say, the last month and a half has been quite a challenge and I think I’ve never seen a project go in any direction like I’ve seen this one going… But I saved Christmas … once again (and the twitterverse won’t be so happy, cause now there’s time again for that: SORRY TWEEPS :-) )

The reason why there was radio silence is that I’ve been battling a ferocious dragon : a 2 headed language beast which consisted out of a 68 projects-large solution (C#and C++). Battling this on the field of both packages, dependencies, but most of all on the field of inconsistent environments and bad TFS installs. Unfortunately.

Why am I writing this down you say? Well there’s 2 reasons: first of all I just wanted to vent all my frustration on a product which I love so much, but this time I had no power of and second of all to give some tricks to people struggling with the same issues somehow (#sharingiscaring is still my motto! )

This entire project is based on a TFS 2010 implementation, for TFS 2012 I get to redo this next year … sigh :-)

Key takeaways from this project:

  • when doing build automation on TFS for combined (or plain) C++ (unmanaged and managed :-( ) and C# solutions keep in mind following TFS system requirements:
    • install the build system in following order
      • TFS Build controller / agent with the lastest SP and fixes / patches
      • install Visual Studio 2010 with all languages (elseway C++ WILL NOT BUILD)
      • install the Windows SDK
      • install Visual Studio 2010 SP’s
      • NOTE : whenever having following issue :  “MSB4019: The imported project “(DRIVE)(PATH)Microsoft.Cl.Common.props” was not found.” when the drive and path are pointing to a different folder tahn the one where all .NET an VS stuff is installed –> abandon the server and redo your installation, for mostly things CPP are missing
    • Native building (C++)  : try to have a 32bit system only as a host when building for 32 bit systems, it simplifies life a lot (i know it’s ugly …)
    • when doing combined building, make sure you do a good amount of tweaking in your build configuration management)
      • have your C++ build with Win32
      • and have your C# build with x86
      • this applies when doing builds for unmanaged C++ only (my guess is that the managed part is still “managed well” by TFS
    • reason is that unmanaged C++ doesn’t cope to well with TFS hosted on x64 environments apparently (it can not differentiate 32bit sources from 64bit sources)

so with this done, i can go and enjoy the Holidays!

Merry Christmas and a APP-y New Year

TFS Preview, Error403 and #windowsazure Continuous Integration/Delivery

Last week I delivered a 1 day bootcamp for Microsoft on the Windows Azure Platform. One of the killer demo’s I prepared was the one on integrating TFS or GIT as Continuous Delivery platform. This being a fairly straightforward demo, I still take the effort of going through it a couple of time jus to make sure that nothing has changed on the platform (as any good cloud platform should also have continuous delivery). And luckily I did …

I had prepared a small MVC demo which I wanted to deploy during my demo. I tested it the evening before and all went fine. As pessimistic I can be I took the non-risk of retesting it again a couple of hours before I went Live … and guess what : all of a sudden my demo would work anymore … Even better none of my projects would build anymore. And for once I was sure I didn’t change anything …

First thing I thought of was: an update must have been pushed to the TFSPreview platform. So I tried to create a new project collection and push from scratch. Unfortunately it gave the same result.

But what was the error … it seems a 403 : Forbidden status was send from TFS, but why? It states me that my build couldn’t access certain folder of which you don’t have any control at all …. Pretty weird.

Now, being a little desperate I dropped the demo of TFS, and let it simmer for a couple of dlays. Yesterday I reached out to the Twitterverse, to see if anyone else is experiencing the same issue, because Binging the issue didn’t really gimme any results.

Luckily my tweet was picked by Buck Hodges (@tfsbuck) a TFS expert and Software Development Manager at Microsoft. He pointed me out to an article on TFS-Azure forum. And that explained it all. It seemed that my certificate had expired (the one created on azure and used for authorizing the auto build and deployment) and therefor the security token wasn’t valid anymore, thus revoking my permissions on TFS for executing Builds.

To solve it is a little more hassle than you want: You actually need to disconnect every Windows azure connection from your TFSPreview environment. Why all of them (well actually only those who are linked to that subscription on Windows Azure which is impacted)? Because the certificate is shared over all the projects.

How do you do this?

  1. Fire up you portal in a browser of choice
  2. From the main screen, press the settings icon in the right upper corner:

  3. Select one of the impacted projects and choose “View the project administration page”
  4. Once in there select the tab “Services.

    Choose disconnect, this will “unlink” both environments on the TFS Side

  5. Then fire up your Windows Azure Management Portal, go to the impacted site dashboard and select “Disconnect from TFS” there:

  6. All you have to do now is to select “Set up TFS Publishing”

    fill out your env name and click “Authorize Now”

    Accept the Permissions settings from TFSPreview:

Select the Project you whish to reconnect

And automatically the Environments will be linked again on boths sides and the indication will be made which version was last deployed (it keeps an history of it J)

And that’s it.

Nothing more to it. It will re-create the cert for you and allows you to deliver continuously (if you have stable builds of course J )

Exchange Server not available …. W00T?

Experienced something funny and i didn’t even realize what caused the issue after 2 days … I spend a lot of time at the customers and my Outlook is my most precious asset for having proximity with my HQ and collegues (ok, ok call me old fashioned, but then again not everybody is connected to Lync yet at HQ. But we’re working on that :-) )

All of a sudden on Monday my Outlook 2010 client started prompting for my credentials. I first thought there was a connectivity issue (the Wi-Fi connection isn’t that stable at the customer, unfortunately. And unfortunately it’s the only for me to connect to the net). Even funnier: all my MSN contacts started appearing offline.

So first thing you try: Reboot. Dind’t work. Second thing you’re gonna try … reboot again. If that still doesn’t help, you start throwing down your defenses (e.g. Firewall and AV). No success .. :-(

Comming down a little frustrated, I tried following reaply some patches, then repaired Office, re-apply SP1 all without success.

As you can see I was getting quite desperate. What the hell was going on and why so suddenly?

Today, after 2 days of trying and “fooling” around I tried to conncet through VPN to see if I was able to fix it when i was “on-premise” or domain connected. So i juiced up the VPN connection and then it stroke me: the VPN solution warned me of not having the latest AV definitions. Strange because all seemed green … so i thought …

taking a look at the definition date i noticed it was dated last friday of the last update but the definition was dated today. W00t?!? checking the date in the system tray explained everything now to me … the date was 5 days beyond the current date, instead of 23rd of May it stated 18th May.

Changed it back to the real date and all of a sudden everything seemed to work normal again. Figures of course, knowing that Exchange, AD and MSN don’t like out of sync devices :-)

But then there’s the matter: what caused the date chance? My guess, someone on the guest Wi-Fi juiced up an AD with Win32 time services and started broadcasting the wrong date, but then again I can’t be sure about that.

BTW: on my quest for a solution I also consulted the big bad internet and most of the time my queries ended up in this kind of articles like this one and this one (searched for ” “outlook 2010″ credentials prompt ” on (don’t like the other one , but that’s another story :-) )

So to add up: whenever having conncetion issues: check your sys date, you’ll never know you might be out of sync!