This weekend a (great) team of engineers from our x86 virtualization group upgraded two of our ESX farms to the latest service pack. Unremarkable in itself until you step back and look at what occurred and what did not.
In a physical world it is rare to be able to patch a BIOS without a hard or soft reboot, it is highly unlikely to accomplish this without some down time on a single instance server. In our VM world we were able to move load off the servers during the upgrade period, redistributing it across the farm, we then were able to apply the patches, update the server and bring them back on line with zero down time to the associated VMs. Okay so no big deal right we updated about a dozen physical servers in the two farms, but think about what occurred at the VM level, in excess of 200 servers received patches with no down time and no issues. This is a truly new experience for us; uninterrupted infrastructure upgrades, and one of the many unintended great consequences of server virtualization.
0 comments:
Post a Comment