“Paging Doctor Dwayne,your stress test patient has arrived”.

For many years I heard on the internal paging system Doctors being summed to baseline their patents, to find out where issues may be lurking. In a pervious VDI tip I mentioned that you need to properly stress test your storage system. This still needs to happen to get a proper baseline and will help you with trouble shooting down the road. This does not exclude you from doing a proper environmental stress test. As you scale you need to make all the parts are working together to form one solid application called “VDI”.

With talking to Mark Plettenberg from LoginVSI, he mentioned a customer that went into production with only 1 CPU running on their storage array. A POC wouldn’t have helped out in this situation. When numbers hit the 400 users mark, performance issues start happening and VDI looks bad to your end users (your customers). It can also help find other configuration issues like your anti-virus causing negative impact with scanning more than it should.

If you’re a VMware View shop you can use VMware View Planner or LoginVSI. LoginVSI is vendor neutral so lots of consulting companies lean towards LoginVSI so they can do both View and XenDesktop. Both are able to launch third party applications to get a real test of your environment.

In the VDI game consistency and reliability go a long way. Don’t get caught off guard. Stress test your environment.

Client-side caching provides huge bandwidth savings with VMware View 5. Client-side caching can save 30-40 percent of the overall bandwidth. The default size is set to 250 MB and can go up to 300 MB. With older thin clients that have less the 1 GB of RAM you may run into issues. If the client-side cache is set too high, you may start experiencing dropped sessions.

Thanks to Chuck Hirstius, the Master of PCoIP at VMware for making me aware of this issue. Give him a follow at @rexremus

Do you want a wireless PCoIP Monitor with a super long battery life and in a small form factor?

Cisco-Linksys WET11 Wireless Ethernet Bridge,a Samsung NC420, ergotron cart, plus a small APC UPS on the bottom of the cart forms a happy solution for medical staff. Also the lack of sound make it great for operating rooms and patient areas. Any wireless bridge will do, I have even used apple airports before. The only requirement is the Samsung NC240 because of the two network ports that come with the unit. The whole solution will run under 100 watts.

The below picture was taken from View Best Practices White Paper but what does it all mean?

pae-SVICreationRampFactor and pae-VCAIIRefitRampFactory are used to set the rates for provisioning with the VMware View composer database. I suspect VMware set it low to acomdate users that install the software on old test equipment and wouldn’t be able to handle the load. You wouldn’t want to lose a proof of concept now would you?:-)

While pae-VCAIIRefitRampFactory still remains a mystery to me, I was able to track down some information on pae-SVICreationRampFactor.The maximum value for pae-SVICreationRampFactor is 50. If this value is set to be larger than 50, the value 50 will be used. Adjust this number cautiously and not ramp it up to high right out of the box. For my testing I left it at the recommend values. When I was composing 10 machines with the default values vs new values I saw a decrease of 2 minutes. If you look at the chart above it looks like 10 machines comes close to where the default settings had a spike. While 2 minutes with 10 machines is not that big of a savings, spinning up a DR would have lots of value with this setting. It’s also interesting that pae-SVICreationRampFactor also speeds up the deletion of machines.

If you’ve ever have had to do maintenance work on a weekend you will appreciate this setting. I will try to continue to dig for pae-VCAIIRefitRampFactory information but it seems to be an uphill battle.

For informatkion on how to backup and connect to the ADAM database click here.

You can change the settings by going here:

Thanks goes out to my TAM Shawn Bertin for helping me out late at night to find out this information.

Backing up all of the View components at the same time might seem like an obvious thing to do but it can be harder than you think. Failing to get close to the same backup window for the View Composer DB(SQL) as the ADAM database can leave you in a bit of a mess. Records may not matching up on a restore will have you deleting virtual machines or records from other parts of the ADAM database to fix inconsistencies.

Inside of the View Administrator console there isn’t a set time that you can select from. The Time starts at Midnight and you can only adjust the frequency, hourly, weekly and so on. You can just adjust it to the hourly setting and that should get you close to what your DBA’s are doing.
I don’t like running backups all day if there isn’t a reason . To get an exact start time you have to head off into the ADAM database. Please read this post for help backing up and connecting to the ADAM database before making any changes.

In the ADAM database Expand OU=Properties, select OU=Server, select CN= in the right pane, and choose Action > Properties to modify the pae-LDAPBUTime. The pae-LDAPBUTime property is the Number of minutes after pae-LDAPBUUnits to perform a backup. Since we left the default of Midnight everything will be based off that time. Therefore, if we but in 60 the next backup should fire at 1:00 am.

IT WORKED!

Hope that helps. If you have a thought or questions drop me a line.

Working with the ADAM database in VMware View is kind of like going to the dentist. You never want to go there, your reasons for going are usally related to pain, but your glad when it’s all said and done. This post will form the ground work for a series of tips & articles over the next couple of weeks. The article references all locations in terms of Windows 2008. If you’re running Windows 2003, god bless you.
Step 1 – Always Get a Current Backup
In the VMware View Administrator console you get create an on-demand backup Under View Configuration -> Servers.
Once the backup is complete, go to the server you ran the backup on and move the .LDF and .SVI file to a “safe location”.

The backup location is: C:\ProgramData\VMware\VDM\backups
.LDF = ADAM database
.SVI = View Composer Database

Step 2 – Connecting to the ADAM database

To connect to the View ADAM database:
1. Log in to one of the View Connection Servers.
2. Click Start > Administrative Tools > ADSI Edit.
3. In the console window, right-click ADSI Edit and click Connect to.
4. In the Name field type: View ADAM Database
5. Select Select or type a Distinguished Name or Naming Context.
6. In the field below, type dc=vdi,dc=vmware,dc=int
7. Select Select or type a domain or server.
8. In the field below, type localhost
9. Click OK.
10. Click View ADAM Database [localhost] to expand.
11. Click DC=vdi,dc=vmware,dc=int to expand

NetApp and McAfee are teaming up to deliver enhanced security for Network Based Storage. The product was launched in January 2012 and VirusScan On Board for NetApp is a fully integrated solution. The only catch that I can find is you do need to be running ONTAP 8.1. ONTAP 8.1 is still only available as a release candidate as the time of writing this article.

The management of VirusScan On Board is done through the NetApp management console. I see this being a problem for most customers. Storage and Security usually sit miles apart in the organizationally flow chart. A shining light however is having your security scale alongside your storage. Set and forget is always good in my books.

The timing of finding this product information couldn’t have come at a better time. VMware just released a paper on Antivirus Practices for VMware View 5. In the VMware paper it talks about not scanning the user persona file shares and setting different polices for ThinApp applications. VirusScan On Board has the flexibility to set different on-access and on-demand policies. You would be able to set on-demand scans for the user persona information and set only inbound on-access scan for your ThinApp Repository.

For more information read the article from McAfee.

I am very excited and honoured to be a delegate for the next Tech Field Day. Virtualization Field Day 2 is running from February 22 – 24, 2012 in Silicon Valley. I’ve always thought that the Tech Field days have been great whether you were a delegate or watching the streaming content. While all my expenses are paid for getting to the event and during the event, I am under no obligation to blog about anything I see. Will I blog about the presenting sponsors? Most likely because I like technology and sharing what I know. I am sure the content will be great as past events.

Two of the presenters have already be mentioned on Twitter at @TechFieldDay . One of those presenters I have a fair bit of experience with their product. I hope that presnter will give us a road map on their up and coming releases. The other presenter I really don’t know much about at all, so it be interesting to see what they have to offer.

Once the official Presenter list is posted make sure you book some time off to watch it live via the live stream. You’re always welcome to ask questions or if you get tied up at work you can download the content later. The content is usally available about a week or two after the event. If you have something you want me to ask prior to going,let me know. I’ll do my best to get ask the question(s) on your behalf.

A specific vendor you want to see?, Nominate one here.
Want to become a delegate?, learn more here.

I’ll start off by saying the vast majority of users will never be able to tell the difference with Build to Lossess(BTL) turned off or on. VMware recommends in its PCOIP Performance Best Practices to have it turned off by default. BTL turned off is one of three major bandwidth saving in View 5.0. The other bandwidth saving features are the client side caching and the lossess codec for text.

I can see the need for it with special uses cases like in the Medical Field or high end graphical workstations in the office, but when you go home at night you just want to check an couple of emails. So lets begin.

All the PCOIP sessions variables can be controlled by Group Policy Administrative (ADM) Template files. The View ADM template files are installed in the install_directory\VMware\VMware View\Server\extras\GroupPolicyFiles . To get this to work we need to use:

vdm_agent.adm – used to run a viscual basic script or PowerShell script on Connect or Reconnect

pcoip.adm – used to set the defaults for Frame Rate, BTL, Image quality and many more.

The PCOIP will take the settings you provide and apply them at the time you connect to your virtual desktop. If you want to verify the settings are being used you can check the logs on the VM at :\ProgramData\VMware\VDM\logs ( I am pretending everyone is running Windows 7). The first few lines of code in the pcoip_server*timestampOfConnection*.log. will show the settings, see below.

02/02/2012, 22:04:16.737> LVL:0 RC: 0 MGMT_ENV :cTERA_MGMT_CFG::load_server_config_from_stores[1]: Did not process over-rideable pcoip defaults from registry.

02/02/2012, 22:04:16.737> LVL:0 RC: 0 MGMT_ENV :cTERA_MGMT_CFG::Registry setting parameter pcoip.enable_build_to_lossless = 0

You can only detect where the user has connected from after the fact, you will have to make your users disconnect to make the necessary changes. So I guess it’s not that automatic but it’s the best I good come up with.

How do you tell where the user connected from?

You can find out from the following registry key –
HKEY_CURRENT_USER\Volatile Environment\ViewClient_Broker_URL
If they are connecting from outside your corporate network it should be the name of your security server which you can check for. If your using an F5 device and no security servers you will have to use
HKEY_CURRENT_USER\Volatile Environment\ViewClient_IP_Address
and create your own logic. PowerShell might be the easiest to use because it has a LIKE operator which will make a quicker job of comparing strings.

So here is the code that you use to accomplish the task of turning BTL off. I named the text file Flip-BTL.vbs

'Declare Environment Variables

End If

To get the code to run when the user connects into their session we must use the vdm_agent.adm. The graphic below shows a local policy but you should use an AD one for better manamgment.

It may be hard to believe but not all the prices for an outlook signature are stored in the same place. Out of the box if your using ProfileUnity from Liquidware Labs you need to change or create a new Portability setting. I have chosen to add rules to the Office Potability Rule set.

Registry Rule: Replace Tree software\microsoft\windows nt\currentversion\windows messaging subsystem\profiles
Filesystem Rule: Replace Application Data Microsoft\Signatures