Category: Hyper-V


Very interesting issue on this to which i am still trying to isolate why it happens.

i rebuilt 15 servers in lab today and decided to script the enablement of hyper-v with start /w ocsetup Microsoft-Hyper-V

i got the prompt to reboot and then off it went installing hyper-v

I went looking for the hyper-v manager icon to import some VM’s but i could not find it!!

i went into server manager and it said the role was not installed … starting to get confused here

so i decided to uninstall hyper-v using the command line…

start /w ocsetup Microsoft-Hyper-V uninstall

reboot the servers as part of the process and re add it with the server manager gui….

and hey presto all better Smile

strange one…..

Advertisements

How to virtualize SBS

This is something i am coming across alot lately and i taught i would share some information which i think is valuable

So first of all yes you should virtualize SBS, it makes alot of sense for multiple different reasons….

So lets start with the basics….

and here is the good news… if you buy SBS Standard you get rights to install Windows 2008 R2 Standard and you are still entitled to the one virtual OSE now that is cool!  if you buy SBS Premium you get 2

if you look at the attached diagram you will see you should install the base sbs on the parent partition and the applications in the virtual environment (after you enable the hyper-v component)

This options up plenty of really interesting oppurtunities to give customers virtualization in a small business that doesnt freak them out as they can see a familiar interface in windows 🙂

Its something we have been doing a long time…. and i think its the best practice way of deploying SBS

I love these blog posts by the Ask The Core team and i think everyone should subscribe to their blog, they do not post often but when they do its worth while….

 

this post represents a problem for which i have seen many a time what it comes to clustering with hyper-v and let me assure you to date its never been a windows / hyper-v issue, it has always been network related / fibre channel related.

In my experience it is almost always down to a dodgy cable but in the case your problem is not the same this guide will help you on your way to discovering the exact issues.

 

Enjoy and thanks to the Ask the Core Team Again!

http://blogs.technet.com/b/askcore/archive/2010/12/16/troubleshooting-redirected-access-on-a-cluster-shared-volume-csv.aspx

 

I had a problem last night which was not necessary the type of problem you want on a friday night….

I got  report that the one of the volumes of a cluster array had failed.

So i did the usual checking, logged on to the san checked the volume was on line and it was… so the basic sanity check… 🙂 and it was registering used space…. another basic check…

Second i logged onto each hyper-v node and checked it’s iscsi initiator was running and connecting to the IQN for the lun that was missing

it wasnt so did the simple task of connecting the iscsi initiator to the appropriate IQN

But the lun could come back online in the cluster manager.

We are now into dangerous territory… and have to be very careful…

covered the basics and disk still not online….

Ok so on the event log….. check it at it said the expect disk signature was different to the one it was expecting so it could mount the disk…

I logged on to each hyper-v node ran disk part, selected the disk i wanted to check its disk signature on and ran detail disk (which gives you the disk signature)… and VERY STRANGE the disk signature was correct to what it was expect on all four nodes…

so the event log was a bit of a red herring…..

I decided it would be best to remove the disk from cluster shared volumes and as a cluster resource… disconnect the iscsi initiator on all but 1 node and try to force the volume online (i had a backup of the volume so i wasnt too worried but it can be a risky step)

and another strange message… it said the volume was right protected… RIGHT PROTECTED! i taught how the heck did it get that… but apparently it is by design ….

so how to get right protected off….

again using diskpart and selecting the disk you want to use

type detail disk and look for the read only setting to ensure it is actually set (current read-only state = yes)

image

once you confirm it type attribute disk clear readonly

and hey presto! the volume is online  and the data looks good…

all you got to do is add it back in to the cluster and mount the volume and bring the VM’s online…