SQL 2005 Express Configuration Guide for SRM

It’s often the case that we as IT engineers must work with several different technologies to configure an overall solution.  A prime example is Site Recovery Manager from VMware.  SRM requires skill sets mainly in virtualization and storage, but other peripheral requirements such as SQL for the configuration database broaden the expertise needed to setup the entire environment.

While you can certainly muddle your way through the installation of SQL and creation of a DB, it’s better in my opinion to know without doubt that it is setup correctly the first time.  It’s even better when you can do so without having to pull in a DBA.  That’s where the following guide from the VIOPS site proves handy.  Rob Nourse of VMware Canada wrote this great document which provides step by step instructions on how to install and configure SQL 2005 Express Edition for use with Site Recovery Manager v4.

http://viops.vmware.com/home/docs/DOC-1611 

XenApp Sessions Cleaner 4.0.2

Listed below is an interesting, free tool that I found on www.citrixtools.net which should help with day to day management of most XenApp environments.

http://www.citrixtools.net/en/Home/newsid374/386.aspx

Disaster Recovery: It’s Really that Easy

So how many of you out there have been involved in or are responsible for ensuring that your company’s critical systems always remain online?  It’s a daunting task sometimes, who am I kidding almost all the time, but take heart as there are alternatives to a career of manual recovery exercises.  Even better, these options build on top of virtualization, a technology most of you should already be using to reduce resource utilization and management.

Bottom line, if you have VMware ESX in your environment along with replicated storage then you can layer on Site Recovery Manager to automate the failover and recovery of your virtual environment, from servers to workstations.  And recently SRM was upgraded to include support for vSphere, ESX 4.0, and NFS storage. 

The following 2 documents from VIOPS outline leveraging Site Recovery Manager in conjunction with EMC storage and VMware vSphere to configure fully automated failover and recovery of virtual systems.  So whether you have a FC based SAN or are using NFS for shared storage connectivity, you should be leveraging SRM to automate DR in your environment.

http://viops.vmware.com/home/docs/DOC-1602 - Steps to setup SRM 4.0 with EMC Celerra NAS Replication

http://viops.vmware.com/home/docs/DOC-1227 - Steps to setup SRM 4.0 with EMC Clariion Storage

NetScaler VPX: Tech Preview Available for ESX 4.0

So as promised in my previous post on NetScaler VPX, the virtual edition of the Citrix network appliance solution has been released for ESX 4.0.  Actually, I found the official release to be kind of low key so therefore I have provided all the relevant links below in case you missed it.  While this is currently a tech preview it will at least allow for VMware environments to begin leveraging this technology in preparation for the production release.

https://secureportal.citrix.com/MyCitrix/login/EvalLand.aspx?downloadid=1857216&LandingFrom=1002 - NetScaler VPX Download Site

http://support.citrix.com/article/ctx123096 - NetScaler VPX for ESX Installation Guide

http://support.citrix.com/article/ctx121291 - NetScaler VPX 9.1 FAQ

http://support.citrix.com/article/ctx122426 - NetScaler VPX Licensing Guide

http://support.citrix.com/article/CTX121819 - NetScaler VPX 9.1 Getting Started Guide

http://support.citrix.com/product/nsad/vpx9.1/ - NetScaler VPX 9.1 Knowledge Center

VMware View: How to Un-associate a User from a Desktop

During a View implementation this week a customer asked the question of how to un-associate a user from a desktop if they were using an automated pool of persistent workstations.  This would be required if you had a pool of assigned workstations, meaning that once a user logged into a session the VM they were connected to is always mapped to them after that point.  Then, for example, say a user left the company and you want to remove their workstation from the pool so it can either be re-allocated or destroyed, this is how to accomplish that task.

Using the vdmadmin tool execute the following command from the View Manager server while logged in as an admin:

vdmadmin.exe -L -d poolName -m machineName -u UserName

Alternatively, if you want to assign a desktop to a user, meaning that the first time they go to log into View they receive a specific, pre-created VM, then use the following command:

vdmadmin.exe -L -d <pool_name> -m <vm_name> -u <full_user_name>

Where <pool_name> is the View Manager pool name, <vm_name> is the virtual machine name as seen in VirtualCenter (vCenter), and <full_user_name> is the user name formatted as DOMAIN\username .

Below is a usage sample for this command:

vdmadmin.exe -L -d MyPool -m WinXP-VM -u MYDOMAIN\administrator

 The vdmadmin.exe tool is located in the “%ProgramFiles%\VMware\VMware VDM\Server\bin” directory and is accessible from the command prompt.

SRM 4.0 and EMC Celerra: How to Guide

As I’m sure you have all read by now, version 4.0 of Site Recovery Manager is hot off the assembly line from VMware with some exciting new features.  Probably the most interesting new functionality to me is support for NFS.

When poking around for some technical tidbits I happened upon the following guide courtesy of the VIOPS site.  Authored by Cormac Hogan from VMware, the guide is a step by step whitepaper which outlines how to configure SRM 4.0 in conjunction with EMC Celerra, with a specific focus on NFS based storage.  Even cooler is the fact that you can leverage the Celerra Simulator, available free of charge from Powerlink, along with ESX to get your feet wet with testing this new functionality without touching your production environment.  Ain’t virtualization grand!

 http://viops.vmware.com/home/servlet/JiveServlet/download/1602-1-8390/EMC%20Celerra%20NFS%20CLI%20&%20SRM%204.0%20Setup%20Guide.pdf

Oracle: VMware and EMC vs. RAC

Over the past couple of months I have been working on a very interesting project focused on helping a customer move away from an outsourced model and bring their IT infrastructure in-house.  Their business revolves around Oracle on the backend, so it is very important that, as they move to hosting it internal, the environment is configured with cost, performance, and availability in mind.

Their initial plans included a migration to Oracle RAC.  While there is nothing wrong with this direction, it can tend to be very costly and complex.  RAC provides for the utmost in database availability but can be very difficult to migrate to in heterogeneous Oracle environments.  Upon searching around for some options I found a whitepaper from VMware concerning the benefits of leveraging ESX with Celerra to host Oracle as an alternative to RAC.

Since the customer is moving away from a Solaris specific hardware deployment to an x86 platform for cost saving reasons, it’s kind of a no brainer to continue in the direction of server virtualization.  That combined with the fact that they will be implementing replicated NS-480 arrays gives them a lot of options and flexibility as compared to RAC.  Beyond having a more flexible infrastructure, I was surprised by some of the performance and cost savings outlined in the whitepaper.  According to the document, for comparable deployments using Oracle 10g RAC EE and Oracle 10g SE virtualized, TPS was around 950 vs. 550 and Users were around 19.000 vs. 10,000 both in favor of Oracle on VMware.  On top of the better performance, Software Licensing Cost per TPS was around $2100 for RAC and $250 for the VMware base configuration.

Beyond virtualization the focus on Celerra in the document was mainly around demonstrating the advantages of leveraging multiple protocols for storage connectivity through the different approaches for deployment.  The Celerra platform is key for this reason as it supports NFS, iSCSI, CIFS, and FC in a single, fully redundant, and easy to manage platform.

So if your considering the deployment of RAC in your environment, take a hard look at the following whitepaper.  While the comparison isn’t completely apples to apples, for most environments the virtualization approach can meet the availability and performance requirements at a lower price point.

http://www.vmware.com/files/pdf/partners/oracle/vmw-oracle-virtualizing-oracle-db10g11g-vmware-on-infrastructure.pdf

NetScaler VPX: Ready for Prime Time

Well NetScaler VPX is now GA, actually I’m a little late as it was announced last week.  If you recall, I touched on this subject a few weeks ago when harping on the importance of a true load balancing solution for high availability in a XenApp environment.  As a brief recap, the VPX line of NetScaler is a virtual appliance solution that gives you all of the functionality of the hardware based version but runs on top of a hypervisor, mainly XenServer at this point. 

But take heart VMware lovers, word on the street is NetScaler VPX will be released for vSphere by November of this year.  According to the NetScaler VPX site  a free Express Edition is available for download.  Upon further inspection of the license guide, link below, the Express edition allows unlimited use of the Standard Edition functionality at 1Mbps full-duplex throughput.

http://support.citrix.com/article/CTX122426 

While that isn’t exactly practical for full production use, it does allow you to adopt and gain experience with this technology in your environment without being confined to an evaluation period.  The Express Edition binaries can be downloaded from the following link:

https://www.citrix.com/English/ss/downloads/details.asp?downloadId=1857216

Once your ready to move it to production you have a few choices regarding full purchased versions.  The following is a breakdown of the different VPX appliance flavors along with retail pricing:

	VPX 1000	VPX 200	VPX 10
HTTP Throughput	1 Gbps	200 Mbps	10 Mbps
New SSL requests/second	500	500	500
SSL Encrypted Throughput	1 Gbps	200 Mbps	10 Mbps
HTTP Compression Throughput	750 Mbps	200 Mbps	10 Mbps

VPX 1000 Standard = $15K

VPX 200 Standard = $5K

VPX 10 Standard = $1,200

The Standard edition gets you all the critical functionality around load balancing, short of GSLB which I’m not happy about, and Access Gateway Enterprise edition.  There are also Enterprise and Platinum editions of the appliances, an outline of the features can be accessed here:

http://www.citrix.com/%2Fsite%2Fresources%2Fdynamic%2Fsalesdocs%2FDataSheet_NetScaler_Sept09.pdf

Once you’ve decided which version you will be implementing you can use the following guide to get started:

http://support.citrix.com/servlet/KbServlet/download/20331-102-577426/NS-VPXGettingStarted-Guide.pdf

XenApp 5.0 FR2: Automated Management

A classic stereotype for Citrix XenApp as a technology, in about any IT organization, is that it requires a lot of effort to manage.  This especially holds true in larger environments with a lot of concurrent users, therefore usually translating to a number of servers to maintain. 

While there are a lot of creative ways to add efficiencies into the management of XenApp environments, Citrix as taken the next step in building these technologies right into the product.  With the release of XenApp 5.0 FR2, several features have been integrated to add automation with provisioning, power and capacity utilization, as well as general management tasks such as allowing business units to add users to published resources.  The post listed  below by Vinny Sosa from Citrix outlines all of these critical new features as well as other interesting benefits of 5.0 FR2:

http://community.citrix.com/display/ocb/2009/09/16/What%27s+new+in+XenApp+5+Feature+Pack+2+in+plain+English

Provisioning Server and XenServer Interoperability

While working on a XenDesktop POC this past week I encountered an issue with communication between Provisioning Server 5.1 and the target device client which promptly caused a wave of deja vue.  Could it be the old retry issue that has been around since PVS 4.5 and XenServer 4.1?

As insinuated above by the old product versions referenced, this has been a problem with virtualizing Provisioning Server target devices on XenServer since Citrix first comingled the 2 products.  Apparently depending on the configuration of the network card in the XenServer host a problem can occur with excessive retries on the target device side.  This translates to slow boot times when streaming a VM from Provisioning Server as well as long image capture times when trying to build a vDisk.  It appears to be caused by TCP Offload parameters being enabled on the physical adapter of the XenServer host.

A workaround for the issue is outlined in the following Citrix Knowledge Center article but basically you can set a registry key on the target device side which disables Offloading.

http://support.citrix.com/article/CTX117491

UPDATE: Greg Bulla, a friend of mine and tenured engineer at Old Dominion, shot me an email after reading this post to share some of his personal experiences on this subject.  Upon building a XenServer based guest and executing a capture for a vDisk, Greg encountered the timeout issue noted in my blog.  However, the fix I posted did not address his particular issue. 

After implementing the fix outlined in Citrix article CTX115658, he was able to correct the timeout problem with his VM.  This appears to be another method of disabling Large Send Offload.  The link for this particular article is listed below.  Thanks Greg!

http://support.citrix.com/article/CTX115658