no Traffic to Internet, though you can ping the gateway

Posted by O.Sommer

Tonight I was facing an interesting problem I did not had before.
I just had reinstalled a Server 2012 Essentials VM and was configuring this for internetaccess to patch and configure it.
Though i had entered a vaild IP and a vaild Gatway I misstyped the Gateway IP to be the same as the local OS IP. In my case Recognizing my error I reopened the IPv4 config dialog and change the IP to which is the gateway IP:


Then I tried accessing the internet for Patches and things, but that failed… hmm, well troubleshooting is my job and so.
I went on pinging the gateway IP, which succeded, then I ping a FQDN, which failed…hmm, ok went to check for DNS settings, all ok.

So went on an pinged one of the google DNS Server which also failed:


Interesting… so how come you get to the gateway IP just fine, but not bejond?
Well that’s usually a matter of the Gateway/Firewall/Router, right?
Ok, so checked the setting of the Gateway multiple times all seemed right… As the Gateway is a TMG I went on to monitor traffic coming from the VM:


Ping to the Gateway IP showed up just as expeted and were allowed.

But any other traffic to the internet, e.g. pinging didn’t even show up in the LOG… interesting…

Then I fired a “route print” on the VM and found something interesting, two default gateways:


WTH??!! Ok, now that’s not right and might cause the trouble…ie
So I tried to reset the gateway using the IP config dialog, which did not work as it always kept the wrong gateway of in the deafult route.

Ok, so the let’s use “route delete” and then reset the correct gateway in IPv4 setting:


and after resetting the correct gateway of in IPv4 setings:


after this pinging the web and DNS resolutions started to work just fine:


Wie man VMs aus SC Virtual Maschine Manager entfernt, die den Status “update failed” or “failed” job status haben

Posted by o.sommer

Im VMM kann es manchmal vorkommen, das eine oder mehrere VMs nach misslungenen Operationen de facto leere Hüllen auf Hyper-V Host zurücklassen. Meist ist dort dann nur die Konfigurations XML Datei der VM vorhanden, die aber manuell nicht gelöscht werden kann, weil sie angeblich im Zugriff wäre.

Manchmal hilft einfach ein Neustart des Clusternodes der angeblich die “failed” VM Reste besitzt.

Hinweise darauf wie man die direkt aus der VMM DB löschen könnte finden sich hier:

Folgende Lösung fand ich in diesem Forums Thread:

Technet Referenz mit dem SQL Script in dem man den ObjectState anpassen kann (220 im Beispiel, 107 für die angesprochenen Problemfälle):
auch in diesen KB Artikel referenziert:;EN-US;983839