"Hypothetical" Datacenter Overheating
sronan at ronan-online.com
sronan at ronan-online.com
Tue Jan 16 07:41:41 UTC 2024
Good thing there are no windows at this “hypothetical” location :)
> On Jan 16, 2024, at 1:51 AM, bzs at theworld.com wrote:
>
>
> Something worth a thought is that as much as devices don't like being
> too hot they also don't like to have their temperature change too
> quickly. Parts can expand/shrink variably depending on their
> composition.
>
> A rule of thumb is a few degrees per hour change but YMMV, depends on
> the equipment. Sometimes manufacturer's specs include this.
>
> Throwing open the windows on a winter day to try to rapidly bring the
> room down to a "normal" temperature may do more harm than good.
>
> It might be worthwhile figuring out what is reasonable in advance with
> buy-in rather than in a panic because, from personal experience,
> someone will be screaming in your ear JUST OPEN ALL THE WINDOWS
> WHADDYA STUPID?
>
>> On January 15, 2024 at 09:23 clayton at MNSi.Net (Clayton Zekelman) wrote:
>>
>>
>>
>> At 09:08 AM 2024-01-15, Mike Hammett wrote:
>>> Let's say that hypothetically, a datacenter you're in had a cooling
>>> failure and escalated to an average of 120 degrees before
>>> mitigations started having an effect. What are normal QA procedures
>>> on your behalf? What is the facility likely to be doing?
>>> What should be expected in the aftermath?
>>
>> One would hope they would have had disaster recovery plans to bring
>> in outside cold air, and have executed on it quickly, rather than
>> hoping the chillers got repaired.
>>
>> All our owned facilities have large outside air intakes, automatic
>> dampers and air mixing chambers in case of mechanical cooling
>> failure, because cooling systems are often not designed to run well
>> in extreme cold. All of these can be manually run incase of controls
>> failure, but people tell me I'm a little obsessive over backup plans
>> for backup plans.
>>
>> You will start to see premature failure of equipment over the coming
>> weeks/months/years.
>>
>> Coincidentally, we have some gear in a data centre in the Chicago
>> area that is experiencing that sort of issue right now... :-(
>>
>>
>>
>
> --
> -Barry Shein
>
> Software Tool & Die | bzs at TheWorld.com | http://www.TheWorld.com
> Purveyors to the Trade | Voice: +1 617-STD-WRLD | 800-THE-WRLD
> The World: Since 1989 | A Public Information Utility | *oo*
More information about the NANOG
mailing list