> Killawatt is great, but typically does not capture the initial spike > when you > turn the machine on. You should check the manufacturer specs.
I always watch the initial spike as well as the peak load, but I've never had any reason to record the initial spike, for two reasons: - Normally, the initial spike closely matches the peak load. - It's hard to imagine a scenario when all the initial spikes of all the servers are happening at the same time. My machines are configured to stay down after power loss (learned that lesson the hard way). Which means one of the IT guys is physically present, pressing power buttons to bring machines up. All the same, I'm curious, when you say "check the mfgr specs" what are you talking about? Do they publish a measurement for the initial spike? I'm only aware of the supposed max power consumption rating (or max amp draw) that is published. Whenever I compare the actual (measured) peak load of a machine against the whitepaper for that machine ... Typically the whitepaper indicates 2-5x higher power consumption as compared to reality. This is because the rating of the power supply the mfgr installs must be more than sufficient to power the machine, fully populated CPU sockets at the highest clockrate conceivable, all the disks during spinup, all the memory, the optical drive, all the USB ports charging cell phones, everything to the maximum, and then inflated some for good measure. So if you rely on the mfgr specs instead of measuring actual usage, you'll spend more than necessary on excess UPS and power circuits. _______________________________________________ Tech mailing list [email protected] http://lopsa.org/cgi-bin/mailman/listinfo/tech This list provided by the League of Professional System Administrators http://lopsa.org/
