Vancouver Island Health Authority (http://www.viha.ca) employs over 18,000 health care professionals, technicians, and support staff to provide health care and support services to more than 765,000 people on Vancouver Island and the surrounding area. Gerald Liberçan is a Citrix Technology Specialist whose team oversees the XenServer and XenApp infrastructure and ensures that the users are happy and fully productive at all times.
Island Health relies on the Cerner Millennium EHR solution to provide evidence-based decision making for clinicians, reduce medication variances through streamlined administration and dispensing and improve care for common high-risk conditions as well as increase staff efficiency and patient care with real-time bed management, faster access to results and improved collaboration. Cerner Millennium is accessed as a Citrix XenApp published application by over 5000 concurrent sessions from their BYOD mobile devices, workstations on wheels in emergency or other devices.
VIHA’s Cerner Millennium suite is managed as an in-house solution, and required an extensive change and upgrade of the code in order to allow online ordering of medications. The Cerner Millennium Suite is a business critical core application and any downtime would be catastrophic for the organization. As the first phase of the planned upgrade required a new Linux back end server, changing of the code and moving their XenApp 6.5 farm to PVS. The Roll-out was in its final stages when two of the six team members left, which threatened the ability to hit the anticipated roll-out date.
As they went live with the upgrade, a part of the Cerner Millennium Suite named PM Office was stalling. Every 30 minutes or so one of the end users would have the process hang and the application would disappear. This caused the user to try again and again, increasing the overall stress on the entire system’s resources exponentially.
Gerald asked around and was encouraged to look into ControlUp, which is used in many XenApp and XenDesktop organizations. His hope was that it would provide the visibility into the workings of the infrastructure so bottlenecks could be removed before the go-live date. As Gerald recalls, they were running practically blind without ControlUp.
By examining the health of the entire end-to-end infrastructure including XenApp published apps, provisioning servers and the rest of the infrastructure with ControlUp, Gerald and his team saw that when the PM Office process would hang, the CPU would skyrocket, and with the additional attempts of the users to relaunch the application the CPU would reach 100%.
Using ControlUp they were able to kill the hanging processes causing the CPU exhaustion and keep the users operational. Then, using ControlUp’s management actions they were able to replace the faulty DLL simultaneously on all the servers without causing any downtime or interference to the end users.
Gerald leverages ControlUp’s built in script based action platform to develop a nifty script they named “issue collector”. The script can run on any troubled user’s session without any impact on the user and stores diagnostic logs with valuable information to aid troubleshooting and analysis.
PVS caching can be great, until it gets full, then things slow down dramatically. With ControlUp Gerald was able to use Script-based Actions to collect needed information to keep PVS running at full performance. The built-in ControlUp Registry Controller allowed Gerald to change an incorrect registry key on over 185 servers simultaneously– a feat that no other monitoring application can perform.
With Cerner Millennium, Citrix XenApp and ControlUp, Vancouver Island Health Authority has been able to put into place a highly effective solution that gets the most from the IT infrastructure while ensuring the end users always have an excellent experience and remain fully productive, resulting in reduced costs and risk for Island Health while improving patient outcomes for the population they serve. ControlUp provided the visibility and control to allow Gerald to complete the roll-out with the reduced team, keep it running well, and even reduce logon times from 29 seconds to 14-16 seconds.