Pool Health Overview Tab

A few days ago I blogged about the new version of XenServer Workload Balancing 6.5 and highlighted some features that I thought were particularly interesting for XenDesktop and XenApp users to explore.

In my last blog post (which you can read here), I covered how the recent release of a new version of XenServer 6.5 and a new version of WLB adds more functionality, a better user interface, improved auditing functionality and changes in licensing that make it relevant to more users on the XenServer platform including those using Citrix XenDesktop, HDX and XenApp.

Licensing changes mean this product is now available for XenDesktop and XenApp customers for FREE. I’m really excited about this product and since its release have had a few ideas of use cases where I think it could solve some problems I’ve seen in the field, questions a system administrator may need to answer:

Can I use WLB for General Administration and Maintenance Tasks?

Yes you can and some administrators are already doing this, using WLB to control which hosts are involved in migrations, XenDesktop user Siva Mulpuru sometimes posts tips regarding how to use WLB to do this, you can read more on his blog, here.

Can I prove to the budget owner that yes I need a bigger hardware budget?

The historical reports WLB can generate in management ready .pdf format or as raw excel data for further analysis and graphing; Allowing administrators to easily highlight where their farms are reaching capacity and the business case for additional hardware and infrastructure changes.

Which department is using our hardware and network capacity and should they pay for that use?

Many of our enterprise customers are extremely large organisations, some with more than 500000+ employees and numerous divisions and budgets. The chargeback capabilities of WLB enable companies that need to manage internal accounting and budgets to internally invoice based on a number of system resources.

Is there any evidence Tony from Finance actually spends all day watching Golf on YouTube, as we have all long suspected?

WLB is very good at highlighting abnormal VDI resource usage, if Tony really should be using Excel and Outlook most of the time and his CPU and network usage suggests otherwise it can be a useful tool to give Tony a subtle hint!

How can the farm stay stable when someone sends out corporate video to the entire company?

XenApp administrators fairly often see a scenario where their users come in at the start of day and then logon, check their email, go have a coffee, do a bit editing of a word document, then at 11 o’clock half the company logs on to a video intensive management webinar. This can cause a “storm” on the server resource. WLB can help in a number of ways:

  • Alert the administrator to the fact resource has surged
  • Make recommendations as to how a user can move VMs around to mitigate
  • Automate the moving of those VMs around to spread the load as best as possible
  • WLB allows administrators to keep spare capacity on powered off servers saving power and reducing TCO and bring those servers into use as demand rises

It’s worth noting that XenApp users can ensure new users logging on are allocated to VMs with the maximum available capacity based on a number of optional parameters (find out more, here).

Can I save the planet?

The power management capability of WLB allows VMs to be placed maximising server density and allow spare capacity to be put onto servers that will power off. Saving power, cutting electricity bills (and costs) and, of course, helping a little to save the planet by making your farm more environmentally friendly!

Is today normal for MY XenDesktop/Xenapp farm?

Many XenDesktop and XenApp installations have normal daily usage patterns. WLB has some capacity to raise alerts, recommendations and take corrective actions based on historical trends from the individual farm in which it is installed. I recommend reading the section 5.1.1.1.1 on this in the User Guide around the case study describing a normal “Thursday”….

  • For example, if the CPU utilization on the host exceeds the High threshold at 12:02 PM on Thursday, Workload Balancing checks the data utilization at 11:32 AM on Thursday, and at 12:02PM on Wednesday (the previous day). If the CPU utilization is 80.1% at 12:02 PM, but only 50% at 11:32AM, and 78% at 12:32 PM on Wednesday, Workload Balancing does not make a recommendation. This is because the historically averaged utilization is 72.47%, so Workload Balancing assumes the utilization is a temporary spike. However, if the CPU utilization was 78% at 11:32AM, Workload Balancing would make a recommendation since the historically averaged utilization is 80.1%.

 

This type of trend based alert helps admins handle known temporary peaks in demand such as the 9am logon bootstorm. 

How can I better handle staff turnover?

You find a brilliant system administrator, they are trained up and understand your farm and usage well – and then they quit and leave! The history retained by WLB and the ability to configure thresholds to raise alerts can help you train new staff on the specifics of your farm whilst allowing expertly tuned thresholds to be retained that can alert less-experienced staff to problems.

Further Information