VMware vCenter Heartbeat end of availability enter Symantec Cluster Server and ApplicationHA to the rescue

VMware has announced the end of availability for vCenter Heartbeat, (official announcement here. ) well this came as no big surprise really, if you’ve been watching VMware carefully you would of seen that they have been putting a lot of effort into the vCenter VSA, that’s the virtual appliance which replaces the need for a Windows vCenter deployment. VMware have been adding more capabilities to it and its much more stable now which helps to make it stand up to the intense workloads that many organisations have wanted to use it for. But with the move towards an appliance there are a number of questions to ask, just how will VMware solve the availability piece of the appliance, after all its still a single VM which relies on VMwareHA and DRS for resilience and that’s not really enough. What if I still want to use the Windows version of vCenter, what are my options?
Regarding the VSA and its availability, well I guess they must be close to providing some type of redundancy now, whilst I was walking around the expo hall at VMworld way back in 2012 and came across a VMware R&D stand where I got chatting to one of the engineers there, we talked about the appliance and availability in particular, the engineer was kind enough to talk me through some of the developments that they had been working on around this and gave a demo of a failover between two appliances which was interesting although at the time there was still work to do on some of the monitoring aspects. Now all this is OK if you want to virtualize your vCenter and use the VSA, by doing so you will need to make sure that it is fully protected for resilience cause as I’m sure you are aware it is now pretty critical to the operation and management for the virtual environment.
So what if you want to still use the Windows vCenter version after all the vCenter VSA does eat up valuable resources, well If you are going to virtualize the Windows vCenter version then be aware that ApplicationHA has an agent especially for vCenter and will give additional protection beyond that of just  VMware HA and DRS, check out the article here for more info on using ApplicationHA with vCenter.
 

Please follow and like us:

VMware vCloud Director Technical Whitepapers for SAP with EMC VMAX, VNX and ApplicationHA

Over the past couple of months I have been busy working with the solutions architect guys over at EMC, I’ve been assisting with technical whitepapers for high availability solutions with SAP running within VMware vCloud Director environments. Symantec ApplicationHA featured quite heavily within their designs to protect and provide resilience for SAP, Oracle and SQL.
What was cool was seeing the designs being built and then certain scenarios being tested against high availability SLA requirements. We certainly put the solutions through its paces and I am proud to say that ApplicationHA operated as expected without a hitch and ticked all the boxes in the design.
If your building out a vCloud environment and Tier 1 applications like Oracle, SQL and SAP will be running within it then be sure to give the technical whitepapers a read through, they are stacked with some really useful information and have loads of best practices detailed for use in your designs.

EMC Cloud-Enabled Infrastructure for SAP: High Availability and Application Mobility VMAX

https://community.emc.com/docs/DOC-29575

EMC Cloud-Enabled Infrastructure for SAP: High Availability and Application Mobility VNX

https://community.emc.com/docs/DOC-31655

Please follow and like us:

Comparing VMware vSphere App HA with Symantec ApplicationHA

So before I jump into this post I want to put this disclaimer out there just to be clear.
The following post is a reflection on the views of myself and not that of Symantec Corp.
The reason why I wanted to create this post was that I was reading a whole stack of posts from other bloggers who were mentioning certain things that were just not true, comments such as App HA can protect Oracle, SAP or any application & App HA has been around since 5.0 are a couple of comments I have read on other sites and are just are not true. So what I wanted to do was to produce an honest account of both technologies and highlight gaps and strengths in both solution.
So here we go…..
Solution Functionality
VMware vSphere App HA
The vSphere App HA solution provides the administrator the ability to define high availability for applications running in the virtual environment, and it provides a level of visibility and control via the Web Client Interface for vSphere.  The App HA solution provides levels of remediation by restarting failed components on the application or using the Application Awareness API via VMware HA to reset the VM if the restarting of the application does not complete correctly. The vSphere App HA solution provides support for SQL, Tomcat, Apache, TC Server and IIS. It supports a limit of 400 deployed agents and each VM guest uses this Agent to monitor the applications.
Symantec ApplicationHA
The Symantec solution also provides the administrator the ability to define levels of high availability for applications running in the virtual environment, and employs management via a vCenter Plugin. The plugin provides visibility and control to the applications and also enables the operator to control the applications. The Symantec solution uses the heritage of Symantec Cluster Server, powered by Veritas agents to monitor and control the applications, and supports over 23 tier 1 applications such as SQL, Oracle, SAP, MQ, Exchange, Sharepoint and IIS to name a few, and supports homegrown applications that enterprises may be using to drive their businesses. There are 4 main levels of remediation, ApplicationHA can either restart an application’s failed components, and it can reboot the OS if the previous remediation step fails or restart the VM via the Application Awareness API. Finally it can even recover the last know good backup in case of corruption. The ApplicationHA solution has been performance tested to support 2000+ VMs and uses simple wizards for deployment and configuration. The Symantec solution can also leverage the Virtual Business Service from Veritas Operations Manager to control multiple tiers of applications in a start/stop process and employs fault remediation to resolve issues where ever in the stack a fault may occur.
Solution Functionality Comparison
Symantec ApplicationHA uses its knowledge of applications via its Symantec Cluster Server technology and supports those which are critical to enterprises today.  It has been further developed to simplify the installation, configuration and management of applications running in a virtual environment, and it scales to enterprise requirements.
On the other hand, VMware’s vSphere App HA is very much a 1.0 product and requires multiple points of management. The lack of any deployment automation can be a challenge for some environments and with limited support for key tier 1 applications or home grown applications; it could be viewed as not enterprise ready yet.
 Obtaining the Solutions
 VMware vSphere App HA
The VMware solution is available with vSphere Enterprise Plus edition or available as part of the vCloud Suite bundles which contain vSphere Enterprise Plus as part of the suite.
 

Product Features

Standard

Enterprise

Enterprise Plus

vMotion

Storage vMotion

High Availability

Data Protection

Fault Tolerance

vShield Endpoint

vSphere Replication

Hot Add

App HA    

 Fig 1 – Compare editions for VMware vSphere App HA

Symantec ApplicationHA

Symantec ApplicationHA is licensed either by per VM or by CPU tier (physical ESX or vCPU) and can be installed against vSphere Standard, Enterprise and Enterprise Plus. Depending on the scale of deployment the average cost for ApplicationHA to vSphere App HA is around 20% of the additional cost of Enterprise Plus.  Obviously the Enterprise Plus edition contains other features beyond that of vSphere App HA, and ApplicationHA would be an additional cost but would bring enterprise application support which is missing in vSphere App HA.
Whilst App HA may be sufficient for some environments, there will be environments that need the scope of application support that ApplicationHA offers and the levels of resilience that it offers too. Understandably buying a solution that comes bundled with something like vSphere makes sense, but there are occasions when third party solutions make more sense as they have been developed to provide the functionality that customers need and they typically have the speciality knowledge in that field too. Symantec (Veritas) have been providing HA and DR solutions for over a decade and know how applications work.
Backend Infrastructure Requirements
VMware vSphere App HA
The infrastructure requirements for vSphere App HA are provided by deploying the App HA appliance; this is a dedicated VM which is configured with 2 vCPUs and 4GB RAM. The scale limitation of App HA is 400 agents which take the form as Hyperic Agents. For App HA to operate in the environment it must contain the Hyperic Server along with its database, thus requiring another two virtual appliances. So that’s 3 separate VMs required for the solution to operate. The application monitoring framework is provided via the VMware vFabric Hyperic appliance, which for a small environment (50 Agents) requires around 4 vCPUs and 4GB RAM; the recommendation to support the limit of 400 agents would require 4 vCPUs and 8Gb Ram.  The Hyperic Server also requires a Postgres database which is packaged as an OVA supplied by VMware along with the Hyperic Server. This database appliance typically requires 4 vCPUs and 4GB RAM but recommendation is to increase the RAM to 6GB. In all, the vSphere App HA solution will require at least 10 vCPUs and 18GB RAM.

Once the Appliances are deployed the administrator would then configure “VM and Application monitoring” within VMware HA on the clusters where the virtual machines are residing to enable protection with vSphere App HA.

AppHA1 Fig 2 – VMware vSphere App HA Architecture

Symantec ApplicationHA
For the centralized management of availability protected VMs, Symantec ApplicationHA utilizes the Symantec HA Console, this is installed on a Microsoft Windows virtual or physical machine. The HA Console provides management via a vCenter plugin which is registered during the install process. An installation wizard provides the process to deploy the HA Console Server. ApplicationHA can also protect the HA Console Server by providing an agent to make the HA Console Server highly available. Once deployed the administrator would configure “VM and Application monitoring” within VMware HA for the clusters so that ApplicationHA can utilize the Application awareness API for application heart-beating with VMware HA.
The requirements for the HA Console Server is to have a minimum of 2 vCPUs and 2GB RAM for a 1000 VM deployment and 4 vCPU and 4GB RAM for a 2000 VM deployment; this is well under a third of the resources required by App HA and only uses a single VM to be built and maintained.

AppHA3

Fig 3 – Symantec ApplicationHA Management vCenter Plugin

Backend Infrastructure Requirements Comparison
The infrastructure resource requirements for VMware vSphere App HA is more than triple  that of Symantec ApplicationHA, and with multiple VMs required for the appliances it does consume a large amount of resources and has more touch points to manage to provide the functionality. One of the benefits of the vSphere  App HA solution is the ability to be managed via the Web Client; where as  at the moment ApplicationHA only uses the vCenter Plugin for similar management interaction.
Application Agent deployment inside the Virtual Machines
VMware vSphere App HA
The vSphere App HA solution does not provide any VM agent deployment mechanism and relies on the administrator to provide it. The Hyperic Agent packages are supplied as “RPM” or “EXE” binaries for their relevant platforms.
Once the Hyperic Agents have been deployed to the VMs the administrator would then need to start the configuration of the agents.  The configuration of the agents does require some modifications to the agents configuration file “agent.properties” to enable the vSphere App HA alarms in vCenter. The administrator would have to stop the Hyperic agent if it was running, then make changes to the config file and restart the agent. Note this would have to be completed for each agent running on each VM.
With the agent started the updated configuration is the sent to the Hyperic server.

appha4

Fig 4 – Modifying the VM agent.properties in Hyperic

Symantec ApplicationHA
The deployment process for the ApplicationHA VMs is provided via functionality from the HA Console Server. The Plugin provides a mechanism by which the administrator can select any powered on VMs and deploy the agent “Guest Components” to Windows or Linux VMs.  The status of the deployment is tracked via the task panel in the vSphere UI.

 appha5

Fig 5 – Symantec ApplicationHA guest deployment via vCenter

  appha6

Fig 6 – Symantec ApplicationHA guest installation task status via vCenter

 Application Agent deployment inside the Virtual Machines Comparison
Whilst modification of the App HA configuration is a simple process to carry out, be aware that this must be done for each agent that is deployed and it can be time consuming and error prone especially when carried out manually. The Symantec solution requires no configuration changes with respect to its agents that get deployed, and the Symantec solution is able to push out the agents centrally for both Windows and Linux VMs directly from the vSphere UI.  Also note that the alarm functions are setup automatically with its deployment.
Configuration of Remediation Policies
VMware vSphere App HA
With the VMs Hyperic agents configured and information sent to the Hyperic Server, the administrator does need to configure the remediation actions for the application via a Policy in the Web Client.  If alerting is required for audit tracking then this needs to be configured from within the Hyperic Server UI.
From within the vCenter Web Client the administrator would connect the App HA appliance to the Hyperic Server via a configuration section in the plugin panel. Once communication is made the administrator can then create a policy for the application.  The policy can be created for a specific application or it can be created for a collection of the same type of applications, ie:- create a policy for all SQL 2008 R2 applications or create a separate policy for a specific SQL 2008 R2 application. Note that different applications require separate policies.

   appha7 appha7a

Fig 7 – Creating an application remediation policy & specify failure policies

Symantec ApplicationHA
The deployment process for the virtual machines also configures the connection of the VMs to the Console Server via its own Single Sign On process using the credentials supplied during the install process and requiring no user configuration. There are no policies to define with ApplicationHA during the installation process. Remediation settings can be specified if required once the application has been configured.
Configuration of Remediation Polices Comparison
Symantec ApplicationHA reduces the management overhead in configuration by removing the need to have to create policies up front for the applications. The configuration process for the applications contains the remediation policies and often will require no changes to its underlying policy.
Configuring Application Availability
VMware vSphere App HA
To configure applications with vSphere App HA the administrator first needs to add the server to the Hyperic inventory from within the Hyperic web management page. Once the server has been added to the inventory the administrator creates an alert service for the virtual machine so that it will alert vCenter Server if the application elements fault or have issues.
The administrator then would switch back to the vCenter Web UI and assign the remediation policies against the VM based on the application being monitored.

 appha8

Fig 8 – Adding the VM into Hyperic Inventory

Symantec ApplicationHA
From within the Symantec High availability tab within vCenter the administrator would click on the link to configure the application for high availability. An Application configuration wizard is displayed and the administrator selects which application to configure.  Once selected, the wizard auto discovers the components of the application and displays those for the administrator to select. The wizard then would configure the application for high availability and would display the status of the application for the VM in the Symantec HA plugin tab. Remediation policies are set on a per VM basis from the settings link on the application status view. Some of the ApplicationHA agents utilize detailed monitoring so that the actual health state of the application is monitored.  For example, if a database stopped responding or becomes un-mounted, the agent would react to this and report the application state as faulted.

appha9 Fig 9 – Symantec ApplicationHA application configuration wizard

Configuring Application Availability Comparison
Symantec ApplicationHA removes the need for the user to truly understand the application to be able to configure it. Operations teams with basic application information given from the application owners are able to configure the application quickly and efficiently from a single console. The VMware Solution requires the user to switch to the Hyperic Server UI to add the server to the inventory and then switch back to the Web Client to apply the policy which could mean that different teams handle different jobs based on their access control.
Operating and Managing the Application Availability
VMware vSphere App HA
Operating the vSphere App HA solution is done with the vCenter Web Client. The application status is viewed via the “Application Availability” tab from within “Monitor” at either the Datacenter or Custer level. There is no ability to view the application status at the virtual machine level.  During normal operations the Hyperic agent running on the virtual machine monitors the application state and reacts to failures that may occur within the application. Based on the policy set for the application certain remediation would happen if a fault was to occur. The Hyperic agent carries out service level monitoring of the application and restarts the failed service based on the policy setting. If the failed service does not restart then the remediation policy is run and vSphere App HA solution simply resets the virtual machine. There is no graceful shutdown of the operating system, this is not ideal as possible file-system corruption could occur in such circumstances. Once the virtual machine is reset the monitoring cycle is resumed. Alerting is visible in the Web Client describing the fault and action taken by vSphere App HA.

appha10 Fig 10 – VMware vSphere App HA display

Symantec ApplicationHA
Operating the ApplicationHA solution is done with the Symantec HA tab within vSphere management UI. For environments that limit the user access to the vSphere UI it is also possible to manage the applications directly from a secure webpage, CLI or using the Veritas Operations Manager (VOM) interface which also employs various levels for granularity for access control.
The application status can be viewed at the virtual machine level, Cluster and Datacenter level. The application view also displays the application components in a dependency view so that the administrator can review the hierarchy of the application components. From within all views the administrator can start and stop the application as well as enter maintenance mode for application patching purposes. ApplicationHA provides differing levels of remediation for faulted applications. It can restart failed components of the application a number of times before advancing to the next level of remediation. It can also instruct the VM to carry out an operating system reboot so that the virtual machine is shut down cleanly in an attempt to clear the fault. Finally, ApplicationHA can pass control to VMware HA to reset the virtual machine, or it can also be integrated into Symantec BackupExec to recover from the last known good backup image and restore that as the last step in remediation.

appha11 Fig 11 – Symantec ApplicationHA Management (VM View)

appha12

Fig 12 – Symantec ApplicationHA Dashboard (Cluster/Datacenter View)

 Quick Comparison:  vSphere App HA vs Symantec ApplicationHA

vSphere App HA Symantec Application HA
Architecture Requirements
Management Appliance Multiple, including App HA Appliance, along with the Hyperic Server Appliance and its database. Provided via HA Console Server running inside Windows VM
Management Appliance required for function to operate. Yes, loss of either appliance = loss of protection No, ApplicationHA continues to operate with VMwareHA
Management foot print 10 vCPU 18Gb RAM 1-2 vCPU 2-4Gb RAM
Management UI vSphere Web Client vCenter Plugin
Platform Support vSphere 5.5 vSphere 4.x & 5.x
Installation & Configuration
Push install of Agent/Guest components No Yes via vSphere UI
Management interface points Multiple points of management between Hyperic server and Web Client Single management tab provided via vCenter Plugin
Configuration of Agents Some configuration file modification required Simple wizard to configure application monitoring
Support for custom or generic applications No Yes
Applications Supported SQL, Tomcat, Apache, IIS, vFabric tc SQL, Oracle, SAP, MQ, Weblogic, Siebel, DB2, Exchange, Sharepoint, IIS, Fileshare, Printshare
Number of applications supported 5 with 14 application versions ie IIS 6/7/8 24 with 63 total application versions ie IIS 6/7/8
Management and Operations
Levels of remediation 2 – App restart & VM reset 4 – App restart, OS reboot, VM reset & Backup image recovery via BackupExec
Application Policy Yes, defined per application type or specific application Yes, defined after application configuration.
Application alarms Defined on per VM basis from with Hyperic Server Enabled as default under “Tasks and Events”.
Application control None Yes, Start/Stop
Multiple Application Control None Yes, Virtual Business Service
Maintenance Mode Yes Yes
SRM support No Yes
Backup integration for recovery None Yes

Conclusion
Although on the face of it looks like VMware is providing a solution with similar reputation as VMware HA, their vSphere App HA product is lacking many enterprise features. Two gaping holes are that there is very limited support for any tier 1 applications, and there is no support for custom or generic applications.  Only Microsoft SQL is a real tier one app on their supported list, and there are many major tier 1 applications missing such as SAP, MQ and of course Oracle.
The ease of use capability which VMware often promotes is also missing in this solution. The installation process of the agents is very manual in its approach and requires some type of automation to be provided by the customer. After installation, some amendments need to be made to the configuration along with vCenter alarm configurations which have to be created on a per VM basis before you’ve even configured App HA.  This all adds up to a very lengthy and error prone deployment process, especially if no automation is available.
The levels of remediation are basic, in that App HA can only restart a failed application service and then performs a hard reset against the virtual machine, which is not ideal as it can cause file system or VM corruption, compounding the application downtime even more. ApplicationHA provides OS reboots to safely shutdown the file system and along with integrating with Symantec BackupExec to provide VM image recovery if all goes really wrong. Symantec ApplicationHA works autonomously from its management console and still provides protection via the application awareness API within VMwareHA, ApplicationHA also contains an agent to protect the HA Console server if required too.  The same can’t be said for the App HA solution if the Hyperic Server or App HA appliance were to fail.
VMware does provide management via the Web Client but they do omit any management functions such as starting or stopping the applications. Visibility is only provided at the cluster or datacenter view with no singular VM view or application dependency view. The “Application Availability” view provides very limited information on the application configuration and allows no customization. In a large environment with multiple applications of the same type it will be very easy to become confused with only the virtual machine name to distinguish between application types. Also having to use two management UIs for administration is not ideal and some operational workflow would need to be in place to operate efficiently.
Overall it seems that VMware has hastily put together a solution which seems to be integrated to a point but it is missing some vital operational workflow as well as support for critical tier 1 applications.
Reference Information
Comparing the Editions
http://www.vmware.com/products/vsphere/compare.html
vSphere App HA Overview
http://www.vmware.com/products/vsphere/features/application-HA.html
vSphere App HA Release Notes
http://www.vmware.com/support/appha/doc/releasenotes-appha10.html
vSphere App HA Documentation
http://www.vmware.com/support/pubs/appha-pubs.html
Hyperic Server Configuration requirements
https://www.vmware.com/pdf/vFabric-Hyperic-Supported-Configurations-and-System-Requirements-57.pdf

Please follow and like us:

vSphere 5.5 & 6.0 how to add domain users to SSO

So you’ve installed your ESX servers and installed vCenter along with its SSO, Inventory Services and Web Client, you’ve even installed the Windows vSphere UI just because that’s what you’re used to and now when you’re connecting to your new environment your getting a “You do not have permission to login to this server”
VC Login Issues
So with this new version of vSphere you need to enable access for your domain users/groups so that they have access to vCenter.
To get you up and running fast follow these simple steps. for best practices configuring SSO please reference the VMware documentation.
1. Login to the Web client “https://client-hostname:9443/vsphere-client” with [email protected] using the password of what ever you configured SSO password as. *Note you can only configure SSO using the Web Client*
2. Navigate to vCenter Servers > Manage > Permissions, click on the + to add a user
3. Now add your Domain user or group that you would like to have access to vCenter along with the permissions required and click on OK.
4. Try and login with the user you just configured and you should now have access to the Web client or the Windows vSphere UI.
Adding Domain User to SSO
 
vSphere-land.com is out their to ask your opinion about the Top Virtualization Blogs which have helped you in the Past Year to Learn, Explore & Master the Virtualization Technology & the related Eco-system.
If my Blog has helped you in your journey of Virtualization, I would of course be grateful if you did spend a VOTE for me on their survey.
You can VOTE by clicking on the following link.

Please follow and like us:

The annual VMware & virtualization top blog survey is now open for voting,

Its that time again to have your say on what site should take the prize of top Virtualization Blog of 2013, A big thanks to Eric Siebert at http://vmware-land.com/ who spends quite a lot of time putting this survey together and it’s there for readers of these blogs to vote for their best and most informative blogs out there.
If you would like to cast a vote for this site then thanks and please head over to http://www.surveygizmo.com/s3/1165270/Top-vBlog-2013   and start voting for your best blogs.
Thanks

Please follow and like us:

London VMUG Meeting 24th January 2013 – Register Now

The first VMUG meeting of 2013 kicks off in London on Thursday 24th from 8-30 to 17:00 and the agenda looks well and truely packed to the brim of some cool topics and presentations, if you havent attended this event before then its a great time to start and check out the buzz of this great community.
Registration is live at http://www.vmug.com/e/in/eid=695&source=5
Venue:-
London Chamber of Commerce and Industry
33 Queen Street
London, EC4R 1AP
 
Meeting Agenda

  • Nutanix Presentation – Alan Campbell and Rob Tribe, Nutanix
  • Your Journey to the Post-PC Era – Brian Gammage, VMware
  • EUC Panel with VDI Gurus – Brian Gammage, VMware
  • Deploying vCD 5.1 and VXLAN – Dan Senior
  • VMware HA Deep Dive – Eric Sloof
  • Zerto Presentation – How One Company Averted Disaster from the VMworld 2012 – Joshua Stenhouse,Zerto
  • 10Zig Presentation – Thin Client Technology at Its Peak – James Broughton, 10Zig
  • Big Data for the Uninitiated – Stu Radnidge
  • Update – What’s Here and What’s Coming Soon – Spencer Pitts, VMware EUC
  • VMware Certification Update – Gregg Robertson
  • vCAC – Kim Raynard, VMware
  • Networking and Lunch
  • Onward Drinks at Pavilion End
Please follow and like us:

How to convert existing RDMs to VMDKs without downtime.

One of the requested operations I seem to be getting a lot these days is how to migrate existing RDMs to VMDKs, this has been especially more popular since Veritas Cluster Server now supports VMDKs .  In this two part article I shall take a look on how to achieve this, the first part highlighting how to convert RDMs to VMDKs on a standalone virtual machine with minimal downtime and no data loss using Storage Foundation. The second part will take a look at what steps are needed when Veritas Cluster Server is in the mix.
Let’s take a look at a typical virtual machine that has a RDM attached and walk through the steps needed to migrate the data from RDM to VMDKs
In Fig 1 a virtual machine would typically have RDMs attached to them these would be in either virtual or physical mode.
RDM Fig1

Fig 1 – RDM configuration for a virtual machine

 From within the virtual machine the RDM would be seen as a disk and within the disk there would typically be a number of volumes that would contain the data for the application, the example show is for SQL 2008. A number of mount points had been created within a drive letter for the SQL instance and it is these volumes that will be live migrated to the new VMDKs.
RDM Fig2

Fig 2 – Storage Layout for SQL2008 from within a virtual machine

 On the virtual machine add new VMDKs that will be used for the new placement of the application data.
RDM Fig3

Fig 3 – Adding VMDKs to the configuration of a virtual machine

 Inside the virtual machine notice that the new disks will be seen, perform a rescan or the SCSI bus if required. Once Disk Signatures have been written with the Veritas Enterprise Administration UI the new disks can be added to the Dynamic Disk group which currently holds the volumes for the SQL application, “PROD” in this case.
RDM Fig4

Fig 4 –Adding VMDK disks to the Dynamic Diskgroup configuration for a virtual machine

 Once the new disks are added to the diskgroup then data can be started to live migrate from the RDM disk to the new VMDK disks. This Subdisk move process will leverage a feature of Storage Foundation called SmartMove, this will assist in copying over only the writen blocks on the RDM to the new disks instead of all blocks from the volumes which a traditional block copy method whould use, this SmartMove process will reduce the data migration process especially if the writen blocks are much less then the overall disk size that’s being moved.
RDM Fig5

Fig 5 – New VMDKs ready for volume migration

 The process of moving subdisks with the Veritas Enterprise Administrator is possible to use a drag and drop method for the sub disks from one disk to another but with multiple operations and also for accuracy it’s much easier to do this via CLI and create a script to do this.
RDM Fig6

Fig 6 – Moving a volume subdisk by dragging and dropping

Finding the subdisks that need moving from the CLI can be achieved by using the command vxprint as highlighted below.
C:\Users\administrator.WINDOM>vxprint
Diskgroup = PROD
TY NAME         ASSOC        KSTATE   LENGTH(KB)   PLOFFS   STATE    TUTIL0  PUT
IL0
sd Disk3-01     PROD_MNT-01   ENABLED  204800   0        –        –       –
sd Disk3-02     PROD_DATA-01   ENABLED  2097152  0        –        –       –
sd Disk3-03     PROD_REG-01   ENABLED  1024000  0        –        –       –
sd Disk3-04     PROD_DB-01   ENABLED  2097152  0        –        –       –
sd Disk3-05     PROD_LOG-01   ENABLED  1843200  0        –        –       –
sd Disk3-06     PROD_MNT-01   ENABLED  818610   204800        –        –       –
dm Harddisk1    Disk1        –        4192193  –        –        –       –
dm Harddisk2    Disk2        –        4192193  –        –        –       –
dm Harddisk3    Disk3        –        62652349 –        –        –       –
dg PROD         PROD         –        –        –        –        –       –
v  PROD_DATA    –            ENABLED  2097152  –        ACTIVE   –       –
pl PROD_DATA-01 PROD_DATA    ENABLED  2097152  –        ACTIVE   –       –
v  PROD_DB      –            ENABLED  2097152  –        ACTIVE   –       –
pl PROD_DB-01   PROD_DB      ENABLED  2097152  –        ACTIVE   –       –
v  PROD_LOG     –            ENABLED  1843200  –        ACTIVE   –       –
pl PROD_LOG-01  PROD_LOG     ENABLED  1843200  –        ACTIVE   –       –
v  PROD_MNT     –            ENABLED  1023410  –        ACTIVE   –       –
pl PROD_MNT-01  PROD_MNT     ENABLED  1023410  –        ACTIVE   –       –
v  PROD_REG     –            ENABLED  1024000  –        ACTIVE   –       –
pl PROD_REG-01  PROD_REG     ENABLED  1024000  –        ACTIVE   –       –
The subdisks that need moving are subdisks Disk3-01 to Disk3-06, these will be moved to the new disks, the move operation is carried out with the command vxsd –g(DiskGroup) mv –b Disk03-01 Harddisk1, or vxsd -gPROD mv -b Disk3-01 Harddisk1 in the example shown below.
RDM Fig7

Fig 7 –Movement complete for the first subdisk using SmartMove.

 Notice the –b in the command will carry out the command in the background so that multiple moves can be carried out in one go.
RDM Fig8

Fig 8 – Subdisk volumes being migrated with SmartMove.

 Now that the volumes have been migrated the RDM can be safely removed from the VM, this can be carried out without powering down the VMs too using the hotplug features of vSphere.
RDM Fig9

Fig 9 – Final Storage migration complete

 That’s all the steps needed to move the data from the RDM disks to VMDKs, in the next part of this article we shall take a look at the steps needed when Veritas Cluster Server is in the mix.

Please follow and like us:

vMotion Compliant in guest clustering with Symantec Storage Foundation HA 6.0.x

Today saw the release of Symantec Storage FoundationHA for Windows 6.0.1 , and Storage FoundationHA 6.0.2, which includes additional functionality that enables Windows and Linux within a VMware virtual machine environment. In this era of virtualization one of the main challenges is to achieve 100% virtualization for your existing and new deployments. As customers travel towards this goal they find that providing resilience for key business applications can be difficult to attain and can find resistance from their application owners who are used to the traditional way that resilience is provided with clustering on physical hardware. While it is possible to use clustering inside a virtual machine, doing so can prohibit the use of VMware technology that provides virtual machine transportability between physical ESX hosts, in particular vMotion, Dynamic Resource Scheduler (DRS) and snapshots. Tied to this there is also a requirement that direct access to storage is required via the use of Raw Device Mappings (RDM) which means that VMware admins need to get their storage counterparts to provision storage upfront, doing this means that the VMware admin needs to know how much storage the Application owner needs and typically changing this after the fact can be complicated and disruptive.
So what’s needed is an intelligent way that storage is controlled and accessed between virtual machines and this process needs to be transparent to VMware technologies such as vMotion, DRS and VMware HA. Symantec has been working to provide this functionality and enhance the capabilities of Veritas Cluster Server. Today this functionality is available to preview and test in your environment. Additionally the creation and configuration of the application cluster is simplified and enhanced with VMware in mind. In a matter of five simple steps a cluster is created and the configuration of the application is dynamically discovered to make the whole experience a painless one.
In addition to this our focus has been to leverage the management of VMware environments via the use of vSphere Client to manage various aspects of the application such as start, stop and switchover from one virtual machine to another. The access control is an extension to vCenter administration roles and access can be customized based on your virtual infrastructure.  A pluggable architecture facilitates the use of a browser to access the User Interface if vSphere Client cannot be used in your environment.
Key Application Failover Improvements for VMware in both Windows SFWHA 6.0.1 and Linux SFHA 6.0.2
1)     Same Console Server support for both Veritas Cluster Server (VCS) and ApplicationHA – Users will now be able to manage their virtual environments using ApplicationHA and Veritas Cluster Server using a single vCenter pluggable Symantec High Availability Console.
2)     VCS Storage Support/Integration (in guest, platform) : Provide a way to enable vMotion/DRS in VCS clusters configured with shared storage and deployed on virtual machines in VMware environment.
3)     Application Monitoring and Failover Target Configuration workflow : Using wizard workflow  enable users to Configure application monitoring, Un-configure application monitoring,  Add failover target (to a service group/application), Remove failover target (from a service group/application)
4)     Visibility and Control on using the VCS Cluster View on vSphere Client and the Symantec High Availability Console Dashboard on vSphere Client to show ApplicationHA and VCS application states and overview. The dashboard will provide an overview of the entire data center / cluster from a VCS / AppHA perspective. Users will see a consolidated list of applications running on all the virtual machines
Key Application Failover Improvements for VMware in Windows WxRT 6.0.1
1)     New Application Config Wizard Support for Custom Applications, SAP & Microsoft SQL Server: Users should be able to configure 1 or more SQL nodes for failover in a VMware environment. A documented set of steps will  have to be followed for installing SQL itself before VCS can be configured on these nodes. The User Interface will configure the cluster, SQL service group and also set appropriate values for restarting of the faulted node. It will also configure newer agents required to support a VMware environment
Key Application Failover Improvements for VMware in Linux LxRT 6.0.2
On Linux, we have a release for SFHA 6.0.1.  With the 6.0.2 release we are extending Application Failover between VMware virtual machines.  This  release includes changes in agents to support vMotion and the use of VMDK disks.  This release will support Storage FoundationHA, but no changes will be implemented in Storage Foundation.  Storage Foundation is included to allow customers to upgrade from previous versions.  This product inclusion means that you can upgrade SFHA or VCS from 5.1, 5.1SP1 and 6.0 directly to the latest release.  Additional products, such as Storage Foundation for Oracle RAC and Storage Foundation Cluster File System, are not included in this release as having multiple systems access VMDK disks is not possible with the current agent.
1)     New Application Config Wizard Support for Oracle Databases,  SAP, WebSphere MQ configurations and Generic Applications.  Users should be able to configure multiple nodes for failover in a VMware environment.

  1.         Using Oracle, SAP and Websphere, each node in the environment will  have to be installed with the application binaries and have the application running before VCS can inspect and configure these nodes. The User Interface will configure the cluster, a service group for the specific application and also set appropriate values for restarting of on the faulted node. It will also configure additional agents required to support a VMware environment.  The wizard will configure disks, the application and prompt users for virtual IP addresses to complete the configuration as well as push out the VCS binaries to all nodes in the cluster.
  2.        For applications traditionally configured using the application agent, there is a wizard that walks through the setup of the Generic Application including mount points and virtual IP resources along with cluster setup.

Go grab yourself a trail version download directly from http://www.symantec.com/storage-foundation-high-availability-for-windows and give it a spin today.

Please follow and like us:

What will be new in vSphere 5.1 & Introducing vCloud Suite

With the buzz of VMworld slowly coming to a close with only a few hours left now, I thought I would sit down and reflect on some of the information I have gleaned from VMworld 2012. The big announcement on Monday was the launch of vCloud Suite which allows its customers to create an agile data center, with the intelligence at the software level, extending it to networking, availability and storage, all of which will be contained with a single SKU.

The product is built around vSphere 5.1 and launches at version 5.1 itself. It includes existing products such as vCloud Director and vCloud Networking and Security, and the standard package will be available as a free upgrade to anyone paying an Enterprise Plus license fee for vSphere.
There are three versions to choose from beyond the Standard version vCloud Suite Advanced includes vCloud Operations Management (vCOPS), which automates the process of scanning alerts for significant incidents, making the operations manager’s job easier.
vCloud Suite Enterprise also includes vCenter Site Recovery Manager, which automates the process of recovering a virtual data center from a different site in the event of a disaster.

For more information see vCloud Suite Pricing and Packaging whitepaper
The other announcement which was met with massive applause by the 20,000 attendees at the main keynote was the killing of vRAM licensing models, VMware announced that they will be using a Per CPU model with no limits on the number of cores. Sitting next to me was a reseller who was not pleased with this approach but I guess he was within a very small majority, personally I like the approach as it follows the industry which most are using for pricing metrics.
The new version of vSphere will bring some new functionality and also some faster performance capabilities within the VM. Breaking away replication from Site Recovery Manager was also announced which I expect is to bring inline the functionality which Hyper-V v3 will also have. Looking at what new features are coming down the road although not ground breaking, will definitely attract its SMB and also Enterprise Customers too.

The other announcement which I was expecting as it effects products I look after for Symantec was that of the killing off of the windows vSphere Client, VIX or c# UI which ever you are familiar with. The Web Client will be the defacto standard of interaction with all products by the next major release. It still looks like Plugins will be the bane of this direction though as most vendors will have to re-engineer their plugins to Flex to be able to use the Web Client. The Web client does look good and brings management up to date with its look and feel and will also scale much better than the vSphere Client which can suffer from the White Screen of Death (WSD) if using in a large environment and multiple tasks are happing at once. Scalability will also be better with support for 300+ concurrent administration access too.

So while I start to work on dissecting my attended session reports, I will leave you with some docs which will highlight some of the new functionality coming down the line with the new release of vCloud Suite. Word on the street is that by mid Sept we should see the release go GA, just in time for VMworld EMEA in October.

Please follow and like us:

Countdown to VMworld 2012

Well there’s just less than two weeks to go till the start of VMworld in San Francisco and I’m starting to get excited, I’m looking forward to numerous things, firstly this will be the first time I will be presenting at VMworld US which makes a change to speaking at the EMEA event. I’ve checked out registration for my session “SPO3302 – Gain Control of Application Availability Without Giving up vMotion and DRS” and Im really pleased to see the registration numbers for the session in the multiple hundreds, so far.. It promises to be a good session showcasing new technology along with a technical demo which will change how clustered applications are configured inside VMware environments – more details closer to the event. Register and come to the session on Wednesday 29th 12:30pm-1:30pm
Secondly, I have always read the buzz about how much bigger the US event is compared to EMEA and I can’t wait to start comparing to see if the quality matches, I’ve also managed to swing a full pass so will be catching up on sessions and hands on labs which this year is supporting BYOD which will be fun. I will be blogging about some of these sessions I will be attending as the conference progresses too, so please check back for updates .
And of course not forgetting the networking, it’s always good to catch up with old friends and make new ones along the way too, so I have signed up for a number of parties and tweetups.
Well that’s all for the moment until touchdown in San Fran, watch this space for updates once VMworld kicks off.
If you want to meetup and discuss anything HA related during the event please feel free to ping me on twitter.

Please follow and like us: