I’ve had a couple of occurances of this where a server 2003 R2 vm will seemingly just loose network connectivity. Attempting to ping either in or out will not work. Whilst sometimes powering off and on the vm will resolve the issue it wont stop it randomly happening again (naturally during production hours).
I’ve noted that there is a VMware kb on the issue (which I found after i’d resolved the issue) however it states that it occurs when you have jumbo frames enabled within the guest vm. This however is not the case in my experience.
To resolve this issue just add another vnic to the vm using either the e1000 or vmxnet2 and remove the vmxnet3 vnic the re-enter the vm’s ip information in the newly added card. Another choice would (as a #lonvmug colleague of mine had discovered) be to reinstall vmware tools and reboot the vm Link . Reinstalling the tools is the best bet but I was unable to do that as the vm in question was a terminal services box so I hot added another vnic to avoid rebooting and loosing the disconnected sessions.
There will be some cases where you will need to reboot the vm afterwards such as network sensitive app that cannot handle a short amount of network disruption.
Well, I’m going through my p2v exercise with my current employer and whilst I have p2v’d many servers I had not converted an exchange server before. Our exchange server is a single windows 2003 R2 x64 box with all the exchange roles installed on it, its not a complicated setup but for us it doesn’t need to be. So googling around I found quite a few horror stories of failed p2v attempts. So with a degree of uncertainty I planned for the exchange conversion.
I started off my stopping and disabling these services on my exchange server (because of this obviously you need to carry this out within a maintenance window as exchange will be down for a couple of hours at least.)
MS Exchange Active Directory Topology
MS Exchange File Distribution
MS Exchange Information Store
MS Exchange Mail Submission
MS Exchange mailbox Assistants
MS Exchange replication service
MS Exchange search indexer
MS Exchange service host
MS Exchange system attendant
MS Exchange transport
MS Exchange transport log sear
microsoft search (exchange)
sql server (blackberry)
sql server browser
sql server vss writer
Backup exec
Now there are a couple of services there that are specific to my set up but you get the idea, anything exchange or sql based I stopped and disabled. I then carried out the conversion as per any other p2v using the standalone vmware convertor program. I also wanted to shrink the servers D drive as it was way too big for the mailstore so I configured that as well at this point.
When the conversion had completed I powered down the old physical server and powered on the vm. Then I removed the brand specific drivers and applications (such as HP and dell drivers and array helpers). The I installed the vm guest tools and configured the networking.
I then reset all the disabled services to automatic and restarted the vm. a few nervous minutes later I was able to view emails in owa, use active sync and send/receive emails through outlook, so it all in all seems to have gone well. At least much better than my googling was suggesting it might.
So in my experience of converting exchange server 2007 all you need to do is stop the exchange and sql services (plus any backup agents you may have) and disable them.
Run the conversion using the current vmware standalone converter
Power off the physical
Power on the VM
Remove all of the hardware specific drivers and applications
Install the vm guest tools
Configure the networking
Re-enable the disabled services and reboot the server.
Then check the event logs etc just to make sure windows isn’t complaining about anything.
Then you just need to let your users know that it is safe to use outlook etc again.
The trouble I guess with p2ving exchange is that once you have brought the services back online and it starts servicing mail requests its pretty much impossible to turn the old physical server back on as you will lose any recent transactions that the exchange vm made. So if there is something thats not quite right with it then you really have to power on through and find a fix for it rather than revert to reconverting or powering on the old physical.
Storage profiles in vSphere5 allow you to organise you storage within vSphere based upon its capabilities. vSphere includes some specific API’s that can detect a storage devices specific capabilities however not many storage devices support this function at the moment.
However you can also create user defined storage profiles which allow you to assign capabilities to datastores and then create storage profiles that you can link to the datastores capabilities. This allows you to ensure that your virtual machines get the storage capabilities that they need to perform within your own agreed SLA.
In order to use storage profiles you need to decide what capabilities of your shared storage you want to use to segregate your vm’s with. For example you may have a SAN with a high speed SAS set of disks along with a lower speed SATA array. In this example the choice is quite straight forward as you have a slower capacity based array and a smaller high speed array as well. In this instance you could create storage profiles to represent Capacity or Performance.
With this in mind let’s create a storage profile with these two parameters. Firstly click on Home/vm storage profiles.
1/ Click on manage storage capabilities then click the add button and in this instance lets just call the capabilities Fast and Slow. Create a separate entry for Fast and one for Slow.
2/ Then click on the “Create VM Storage Profile” button. Because we have a fast and slow choice I’m going to use the names Tortoise and Hare as the names of the storage profiles. So give your first storage profile a name in this instance Tortoise then select next. You then need to select the storage capabilities to assign to the storage profile. In our instance we want to assign the Slow capability to the Tortoise Storage Profile. Click Next and then review the information and click Finish.
NB. You can assign two storage capabilities to a storage profile. However because you may also have system defined storage profiles you must not exceed 2 assigned capabilities per datastore. For example you could have 2 capabilities defined in a user defined storage profile but no system defined storage, or you could have one user defined capability in a user defined storage profile and one capability defined in a system defined storage profile.
Create another storage profile for the Hare and assign it to the Fast storage capability.
3/ Then click on the Enable VM Storage Profiles and just click the enable button.
4/ Now all you have to do is assign the user defined storage profiles to your datastores/dataclusters so click on Home\datastores and datastore clusters. Right click on a datastore then select assign user storage capability and select the storage profile you wish to assign to the datastore.
If you have datastore clusters you cannot assign a user defined storage profile to a datastore cluster root however if you assign the same user defined storage profile to each datastore in the cluster then the cluster will inherit the user defined storage profile.
When you are ready to create a new virtual machine you will now have an additional option in the new vm wizard to select the storage profile you need for the vm then it will show you the compliant and non-compliant datastores that you can use on the vm.
Whilst p2ving I came across this little issue when powering on a vm for the first time. I received an error stating the below:
Cannot power On: Reason: The parent virtual
disk has been modified since the child was
created. The content ID of the parent virtual disk
does not match the corresponding parent content
ID in the child.
error
DATE TIME
Power On virtual machine
Now the only thing that I can guess happened here is that vDR created a snapshot of the vm as it was being created. The first thing I tried to resolve the issue was to right click the vm select snapshot and then consolidate. This then thew up another error message similar to the first so that was no good.
I deleted the snapshot and tried again but received again the above error message.
I googled around a bit and found some rather convuluted solutions to the problem however I decided on a much simplier solution that worked just as well.
Before performing any of the below ensure that you have deleted any snapshots that are active on the VM.
1: Right click the vm and select edit settings
2: Select the virtual hard disk and select remove but DO NOT delete the VMDK
3: OK your way out of the settings then go back into the edit settings menu
4: This time select Add and choose Hard Disk and use an existing virtual disk
5: Point your new hard disk at the VMDK file and click OK
6: Power on the VM and all should now be well.
Heres just a quick note for those of you with dell’s sc440 in your home lab I can confirm that the sc440 with Intel Xeon processor IS compatible with ESXI5 so you can upgrade your home lab with esxi5 if you wish.
If you are running a copy of workstation for your vsphere lab and you are encountering an error when trying to enable HA/DRS clusters then this is probably the issue.
When you create a ESX vm, vmware workstation autocreates a vm with 2GB’s of ram which is the lowest amount you can get away with to boot the hypervisor. However in order for HA to work you will need a minimum of 2300megs of ram (2.3GB) so shutdown your ESX/i vm and up the ram by 300megs or more and you will find that HA/DRS works correctly!