Virtualization is all about redundancy. Redundant network switches, frabric switches, power supplies, and of course redundant hosts. Recently, I had to reorganize my entire data center. I was removing equipment and moving equipment. I was trying to reduce everything to either a single rack or a pair of half height racks to save on space, power, and cooling.
Since I was taking out of commission something like 15 hosts, 4 network PDUs, and 8 extraneous switches, I figure the cost savings will be worth the effort. But since I run an enterprise, I need to keep all my VMs running even during the moves of equipment into the new rack(s).
First I VMotioned all my VMs to one host, powered off the host, and physically moved it to the new rack where a new 40 port GBe switch resided. Once that was hooked up and an uplink from that switch to my other network switch fabric was made. I powered on the host and VMotioned all my VMs to the host in the new rack. So far so good. Then the second host was moved and all is still well.
Time to decommission all the other hosts, switches, and PDUs. Doing so freed up a UPS. Granted the new rack had an existing UPS but it is underpowered, so I wanted to move this other UPS into the new rack. All my hosts have redundant power supplies and this was a great help.
I unplugged all my redundant power supplies and then connected them up to the UPS I just placed within the rack. No downtime! I really like this redundancy feature of the hardware. However, I am not done. I like the legs of my power to be even and with this change they were not as both UPS were plugged into the same socket. Now it was time to relocate to another plug that lived on a different circuit and leg of the power box. Once more, redundant power supplies aided greatly.
Now all I had to do was find a place to donate all my old equipment. I found a Technical High School that has a Cisco Academy within it: Keefe Technical High School.
So VMotion tied with redundant power supplies was a way for me to move all my virtualization hosts with zero downtime for the VMs. This is a big big win.
I have lowered my power consumption through four major factors:
- Reducing overall devices plugged in, and removing them from my datacenter
- Choosing power saving equipment
- Upgrading my SAN to something with more space but less overall power requirements.
- Cooling using a new energy efficient unit
Edward L. Haletky, aka Texiwill, is an author, analyst, developer, technologist, and business owner. Edward owns AstroArch Consulting, Inc., providing virtualization, security, network consulting and development and TVP Strategy where he is also an Analyst. Edward is the Moderator and Host of the Virtualization Security Podcast as well as a guru and moderator for the VMware Communities Forums, providing answers to security and configuration questions. Edward is working on new books on Virtualization.