Virtual Infrastructure
-
Fail over Pathing – Since we have moved to the EMC Clariion SAN, we have had two paths to the SAN, but both paths were going through the same switch. The storage team had a change to add a second SAN switch into the storage environment. At the time of this change, the VMware administrators also wanted to immediately take advantage of the redundant switch and move one of the paths to the new switch. This required the ESX hosts to fail over pathing multiple times to avoid downtime. Once completed, we had two paths to the SAN storage going through two separate SAN switches.
-
Co-Located Virtual Infrastructure – We added two customers to our co-located virtual infrastructure. Blair school of music wanted to house a file and print server in a secure virtual environment. So we created a standard virtual machine for them to use for this purpose. The Vanderbilt Institute Research Group or (VIRG) has a database and web server application that the wanted ITS to house in a enterprise fashion. We decided to put the web server into the virtual infrastructure In the process of this addition we put the two new customers on the new firewall protected co-location network, which had to be trunked into the virtual infrastructure with a coordination effort from the Network Design and Engineering team.
-
Service Console Migration – The management interface for the Service Console was on the .115 network, which was not the ideal location for managing the ESX hosts. We moved the service console from this network to the Applications Hosting administration network, which substantially decreases our security risk to this environment.
-
Test Environment – With the purchase of bigger more powerful IBM boxes, we increased the capacity of the virtual infrastructure Also since these boxes have an INTEL chipset rather than our original implementation on AMD chipsets, it gave us an opportunity to re-purpose some of the AMD machines to a ESX test environment. This will be a great advantage when we want to test patching or changes to the virtual Infrastructure.
-
HA and DRS – Once the test environment was in place, it gave us the opportunity to do a proof of concept on VMware High Availability (HA) and Dynamic Resource Scheduling (DRS). This are technologies that make the virtual infrastructure more robust and efficient. HA adds the robust ability to loose a physical ESX host and your virtual machines will start running on available hardware without any manual intervention. DRS provides and efficient environment by making decisions based upon ESX hosts and available resources, essentially distributing the load of the virtual machines evenly across the infrastructure.
Active Projects
-
Helix Streaming Media Server – We have been given the directive to replace the hardware that performs the streaming media services for the university. In this space we decided to host the application on a virtual machine, with NAS attached storage for archive media files. I have built the first virtual instance of the new infrastructure, as we are moving from a one server architecture to a two server redundant architecture.
-
VUNETID Batch, Web, and Database separation – We are looking to tier out the database, batch applications for transactions, and the web portal into the database into three different machines. This is more of an Infrastructure Development project, but we had to provision virtual machines and operating systems for them. We have built two Linux virtual machines so far in this process, and provided them to the iDev team.
-
Boinc Distributed Computing – We were asked to deploy a boinc server so individuals could provide personal computing resources to big research projects that require lots of resources. Boinc has created a pre-packaged virtual machine to act as a server for this process. We downloaded the virtual machine, added it to our virtual infrastructure, and ported code written by the Structural Biology department into the server.
-
Flash Streaming Media Server – Public Affairs wanted to expand the functionality of our streaming media environment to incorporate flash as a deliverable media. We worked diligently to get a virtual machine provisioned and and the application installed and configured to have it ready under tight time constraints. Our streaming media administrator introduced the new service to the community in the web spiders meeting.
-
Load Balancing the Mailgate SMTP environment – We have been working to put the mailgate smtp servers and the front end exchange boxes behind the CSM to balance the load coming to those services. We have come up with a preliminary test plan, and got IP address for this test. We have also pre-configured the CSM for both exchange and the mailgates.