My Virtual Playground

Recently Microsoft have changed the licensing structure of Office 365, and you are now (under certain licenses) able to install Office 2013 Pro Plus on your TS server.  Its not a straightforward clicky the button and install, rather you’ll need a couple of batch files, some changes to the configuration file, and a bit of patience.

• First thing you’ll need is the Office Deployment Tool
• Copy this to a location with plenty of space (about 1.2 GB), running it should extract out the setup.exe file and a configuration file.
• Create a share to that location
• Edit the configuration file to look something like this (where the servername is the name of the server your on, and sharename is the share created above.

<Configuration>
 <Add SourcePath=”\\servername\sharename\” OfficeClientEdition=”32″ >
  <Product ID=”O365ProPlusRetail”>
   <Language ID=”en-us” />
  </Product>
 </Add> <!–  <Updates Enabled=”TRUE” UpdatePath=”default\” />  –>
 <Display Level=”None” AcceptEULA=”TRUE” />
 <Logging Name=”OfficeSetup.txt” Path=”%temp%” />
 <Property Name=”SharedComputerLicensing” Value=”1″ />
</Configuration>

The “SharedComputerLicensing” is the key section, by setting that section we can install and run Office 2013 Pro Plus on the RDP machine and each user will need to activate using their account.

• Next we need a couple of Batch files, one to download the files for a local setup, the second to run the installation

\\servername\sharename\setup.exe /download \\servername\sharename\configuration.xml

\\servername\sharename\setup.exe /configure \\servername\sharename\configuration.xml

• Run both of these in order, and your Office 2013 Pro Plus will be installed.
• Each user will be asked to activate the first time they use the RDP server.

Had a problem with a server which had frozen and needed to be restarted.  When the services all came back up RRAS was showing as available on L2TP, but not on PPTP.  Netstat didn’t show the port as being in use by anything.  Found a few posts that suggested it might be Black Berry Dispatcher, or Router, stopped these services, restarted RRAS, then restarted the Black Berry Services, still no go.  Some other posts make reference to Exchange System Attendant grabbing the ports using RPC which means it won’t show up in netstat.

Stopping the System Attendant and associated processes, restarting RRAS, then restarting System Attendant and associated processes resolved the issue.

I recently configured an SQL Server which we ran in trial to make sure it was what we wanted, after we had tested everything the decision was made to purchase the license. Given tight time frames we were hoping to be able to just get the license key and plug it into the Upgrade Wizard and everything would be sweet. Sadly not the case. We received an OPL from Microsoft, but when you access the licensing portal that doesn’t contain the key because its embedded in the ISO which you can download.

I gave MS Licensing a call and to cut a long story short, they couldn’t provide me with the key and said my only option was to install a new Server. While this was an option it was a long way from my preferred one, due to time and Budget restraints.

I did a bit of reading, and found that to get your CD key if you have an OPL version you can run the installer and it shows you the key when you get to the licence key section of the install. I haven’t tested whether you can do this on the live box without incurring downtime, but I have tested it on another dev box, and I was able to retrieve the key and upgrade successfully.

I’m not sure why the upgrade option is included in the trial if its not a supported option.

UPDATE – Tried to update the key on a Trial machine which had DB’s restored onto it, couldn’t put the key in, got the error please verify the key you are using etc…  However if you mount the ISO the run the upgrade from the ISO it will add the key into the machine, without any down time to the environment that I could see.  I have used this method on two installations and both worked successfully.

I’ve been having some problems with an Exchange 2010 box filling up its disk space very quickly, resulting in an event ID 15006 in the logs.  The log looks like this

Microsoft Exchange Transport is rejecting message submissions because the available disk space has dropped below the configured threshold.

The following resources are under pressure:
Queue database and disk space (“C:\Program Files\Microsoft\Exchange Server\V14\TransportRoles\data\Queue\mail.que”) = 97% [Medium] [Normal=95% Medium=97% High=99%]
Queue database logging disk space (“C:\Program Files\Microsoft\Exchange Server\V14\TransportRoles\data\Queue\”) = 97% [Medium] [Normal=94% Medium=96% High=98%]
Physical memory load = 91% [limit is 94% to start dehydrating messages.]

The following components are disabled due to back pressure:
Inbound mail submission from the Internet
Mail submission from Pickup directory
Mail submission from Replay directory
Content aggregation

The following resources are in normal state:
Version buckets = 0 [Normal] [Normal=80 Medium=120 High=200]
Private bytes = 5% [Normal] [Normal=71% Medium=73% High=75%]
Batch Point = 0 [Normal] [Normal=1000 Medium=2000 High=4000]
Submission Queue = 0 [Normal] [Normal=1000 Medium=2000 High=4000]

This results in no new messages coming into the system from external sources.  Most of the items I read pointed to just clearing up some space, which was done, but within a few days that extra 10GB was all being used, and we are only talking a small environment < 15 people.  This morning the problem was back again, so we gave the vm another 15gb on the fly, and rather than working perfectly in Windows Server 2008, I got a parameter incorrect error when increasing the Disk Space with Disk Management.  We restarted the machine which then showed the correct amount of new disk space, but it had also cleared up another 40GB which I’m assuming was being held in the message logs.  I’ll do some more investigation and try to find why for those who are interested.

If you don’t want to delete anything or can’t add more space to the system, you can move the location of the logs as detailed on the Sinking Ducks, so credit is due there, but I’ll include it here to save you clicking…

• Edit the EdgeTransport.exe.CONFIG file, which if installed to the default location should be “C:\Program Files\Microsoft\Exchange Server\V14\Bin”
• Find the following lines .. and change the path to be the new location you want to use.

<add key="QueueDatabasePath" value="C:\Program Files\Microsoft\Exchange Server\TransportRoles\data\Queue" />
<add key="QueueDatabaseLoggingPath" value="C:\Program Files\Microsoft\Exchange Server\TransportRoles\data\Queue" />

• Create that folder, and restart the Exchange Transport Services

A while back I mentioned to someone at VMWARE about making the documentation available in eReader format, now I’m not going to claim they did this because I asked, but I’m very happy to report that a good portion of the documentation is now available in ePUB and Mobi formats.  I’m not sure when this came about and some probably already know, but I only found it today, and felt obligated to mention it.

The vSphere section is linked below, but some of the other products also have ePUB and Mobi versions.

PDF and E-book Product Documentation

I recently had an issue where the backup software had left a snapshot on the vm, which wasn’t visible in the snapshot manager.  Usually to get rid of this you would just snapshot the machine again, and the process would re-create the missing links.  Deleting the snapshots would then take the machine back to its usual state, happy days.  However this time that didn’t work, I could create snapshots, but couldn’t then commit them, resulting in a few snapshots being overlaid, but not showing up in the snapshot manager.  This was on ESXi 4.1, the backup software was VEEAM 5.

I did a bit of searching online, and found a couple of ways to remove the file locks which were showing up when trying to commit the snapshots.  The first option I tried was simply to disable the backup software which I thought might have the lock on the files.  Because of the situation I was unable to reboot the VM that was running the backup software.  This might have worked, but I don’t know for sure.

Next was to restart the management agents on the host.  Given I wasn’t onsite I wasn’t able to go in through the normal console, and restart the agents, luckily you can do the following.

1. Connect using the vSphere Client
2. Clicky on the Configuration Tab
3. Clicky on the Security Profile link
4. Properties
5. Remote Tech Support (SSH)
6. Options
7. Start

This was originally supposed to be how to update the ESXi ISO to allow for installation by USB to a machine with the Realtek 8111E NIC.  Having spent most of the day trying to get this working either as USB install or Network install, I’ve for the moment put the idea on hold.  Following the instructions available online and using the mkesxiaio script, I was able to install ESXi using USB, but as soon as I tried to use the OEM.TGZ file, the install would get to the reading the Kickstart config file, then erroring indicating autopart –firstdisk –overwritevmfs had no valid target.  Until I can find a solution for this I’ve used the following to get the Realtek enabled.

Things you’ll need –

• Installed ESXi Server with a Realtek NIC
• SSH client (Putty is my choice on Windows)
• oem.tgz file from vm-help

Steps involved –

• Copy the oem.tgz file to a location the esxi machine has access to (NFS share etc)
• Enable SSH access to the ESXi box
• cd to /vmfs/volumes/Hypervisor1
• mv oem.tgz oem.tgz.ORIG
• copy the new oem.tgz file to /vmfs/volumes/Hypervisor1
• Restart the ESXi box.
• You should now be able to add the Realtek to a vswitch

Resources

oem.tgz
mkesxiaio

UPDATE – Something I later found which I now think is the root of my previous problems, when I use the above method to change the oem.tgz file, I lose access to the local datastore after a reboot.  It looks like there is something in the existing oem.tgz that is removed when I add in the new one.  I’m going to try to merge the two oem.tgz files, and see if this resolves the problem.

I know I’m probably a bit behind the times here, given that v5 is already out, but I haven’t done my VCP4 exam yet so I’m still going through the motions for configuring my home lab.  Given I didn’t purchase CD/DVD drives for my lab machines and I haven’t set up a PXE environment yet, though I have plans to do this in the near future, I needed another way to set up the two boxes.  There are lot of examples of how this is done around the interwebs, I tried two methods, both work, though I needed a few modifications to the Windows option. The original information is linked below.

The Linux Option –

For those of you running a derivation of linux the follow instructions worked for me.  NOTE – these instructions are taken from Vallard’s Tech Notes, so credit is due there. 

Things you’ll need –

• ESXi iso, available from VMware
• A USB key (at least 512mb, though I don’t even think you can buy these anymore)
• Linux Box (I used a Ubuntu 10.04 Server)

mkdir /media/ISO
mount -o loop VMware-VMvisor-Installer-4.1.0.update1-348481.x86_64.iso /media/ISO

• Make the USB bootable with a Win95 image.  You will need to check where the USB stick has been attached (mine was /dev/sdb)

fdisk /dev/sdb
d (delete all partitions)
n # new partition
p # primary partition
1 # 1 is the partition number.
1 # the first cylinder
<enter> # the size of the partition, I just used the full disk
a # toggle bootable flag
1 # make partition 1 bootable
t # change the type
1 # of partition 1
b # partition type W95 FAT32
w # write it out

• Next we need to format the disk

mkfs.vfat -n BOOT -F 32 /dev/sdb1

• Using syslinux we make the disk bootable

syslinux -s /dev/sdb1
dd if=/usr/lib/syslinux/mbr.bin of=/dev/sdb # this on in the disk itself (sdb) not the partition (sdb1)

• Mount the USB and copy the ISO contents to the disk

mkdir /media/USB
mount /dev/sdb1 /media/USB
cp /media/ISO/* -R /media/USB

• We don’t need the isolinux stuff…

rm -rf isolinux.bin
mv isolinux.cfg syslinux.cfg

The Windows Option –

For those running Windows the following worked for me.  This is largely taken from VMPros, though there was also some input from others with regards to the errors I received.

Things you’ll need –

• ESXi iso from VMware
• USB memory key
• Windows Box
• UNetBootin
• Syslinux (I got it from Softpedia, but there are lots of other places)

Steps –

• Connect the USB and start the UNetBootin software
• Select Disk Image (towards the bottom), choose ISO, and browse to the location of the ESXi ISO
• Confirm the USB Drive is selected as the destination, and that the correct USB drive is selected.
• Click OK, and if you should receive an error click Yes to overwrite.
• This got me as far as a USB disk that was bootable, however when I tried to install I received an error about  menu.c32 is not a COM32R image.  To get around this I had to copy the menu.c32 and mboot.c32 files from syslinux and replace the existing ones on the disk.  These files are located in the syslinux\com32\menu and \mboot folders.

The rest…

Which ever method you used we now have the problem of the ESXi installer looking for the CD, why… I don’t know.. but it will.  To get around this we need to use a scripted install.  Scripted installs are done using kickstart (originally from linux I’m told).  Though the ks.cfg files between Linux/ESXi aren’t compatible with each other.

First we need to tell the Installer to check the USB disk for the ks.cfg file.   A network location can also be used, but for what we are trying to achieve, we’ll use the USB.  Edit the syslinux.cfg file and add in the ks=usb section of the append vmkboot.gz line shown below.

append vmkboot.gz ks=usb --- vmkernel.gz --- sys.vgz --- cim.vgz --- ienviron.vgz --- install.vgz

And create the kickstart file ks.cfg

vmaccepteula
rootpw YourRootPasswordHere
autopart --firstdisk --overwritevmfs
install usb
network --bootproto=static --ip=<ip here> --gateway=<gateway here> --hostname=ssv-normandy --device=vmnic0 --nameserver=<nameserver here> --netmask=<netmask here>

And yes the host name is after the SSV-Normandy from Mass Effect… Love that game, but change it to whatever floats your boat.

NOTE – THIS WILL OVERWRITE ANYTHING ALREADY ON THE FIRST DISK OF THE MACHINE.

Initially I was hoping to be able to target another USB stick for the installation, however this is not possible in ESXi 4, it is possible in ESXi 5 though.

Resources

Vallard’s Tech Notes
VM Pros
VMware KBs

Having moved back to Australia from the UK, and deciding to start getting serious about virtualization, the only reasonable thing to do was put together a home lab… My Virtual Playground.  After reading about possible set-ups of many others I decided the Shuttle whitebox would be my best option.  The plan is to have a self building test platform which I can throw together from scratch using various PowerShell scripts, PXE installs etc, I’m not sure how successful this will be, but either way I should learn a lot while trying.  As the build comes together I’ll add more and more info on what I have done, and how I’ve done it.  The community for vmware has been so fantastic to me since I started working with vSphere, its time to give something back.

The Purchases –

2x Shuttle SH67H3 bare bones boxes

2x Intel i5-2500s @3.30GHz

4x 8GB DDR3 RAM

2x Intel Dual Port nics (PT, I read that the ET if much better, but for my budget and setup the PT should be more than adequate)

2x 4GB USB Memory Sticks

Netgear GS108T

Thecus 4200eco (support iSCSI, NFS, CIFS, etc)

2x Seagate 1TB HDD’s, raid 5 in the Thecus

Currently the machine are all hooked up, there is still some networking LACP etc to be configured, which I’ll provide more details on later.