There have been some fairly large data set studies done shown that air intake temperature for huge numbers of servers, at 77-78F does not correlate with a statistically significant rate of failure.
http://www.datacenterknowledge.com/archives/2008/09/18/intel-servers-do-fine-with-outside-air/ http://www.datacenterknowledge.com/archives/2012/03/23/too-hot-for-humans-but-google-servers-keep-humming/ how/what you do for cooling is definitely dependent on the load. Designing a colo facility to use a full 208V 30A circuit per cabinet (5.5kW) in a hot/cold air separated configuration is very different than 'normal' older facilities that are one large open room. On Wed, May 11, 2016 at 1:58 PM, Ken Hohhof <[email protected]> wrote: > I’m not sure you can answer the question without knowing the max heat load > per cabinet and how you manage airflow in the cabinets. > > AFAIK it used to be standard practice to keep data centers as cold as > possible without requiring people to wear parkas, but energy efficiency is > a consideration now. > > > *From:* That One Guy /sarcasm <[email protected]> > *Sent:* Wednesday, May 11, 2016 3:51 PM > *To:* [email protected] > *Subject:* Re: [AFMUG] Data center temperatures > > apparently 72 is the the ideal for our noc, i set our thermostat to 60 and > it always gets turned back to 72, so i just say fuck it, I wanted new gear > in the racks anyway > > On Wed, May 11, 2016 at 3:46 PM, Larry Smith <[email protected]> wrote: > >> On Wed May 11 2016 15:37, Josh Luthman wrote: >> > Just curious what the ideal temp is for a data center. Our really nice >> > building that Sprint ditched ranges from 60 to 90F (on a site monitor). >> >> I try to keep my NOC room at about 62F, that puts many of the CPU's >> at 83 to 90F. Many of the bigger places I visit will generally be 55 to >> 60F. >> Loads of computers (data center type) are primarily groupings of little >> heaters... >> >> -- >> Larry Smith >> [email protected] >> > > > > -- > If you only see yourself as part of the team but you don't see your team > as part of yourself you have already failed as part of the team. >
