Re: [AFMUG] Data center temperatures
Absolutely agree… these are not “normal” cabinets for customers … although if you have a cage, you have more freedom to do within it what you want (within reason) hence why these cabinets were chosen/built. >From a DC perspective, yeah the more you can jam in the better (again, within >reason) ;) From: Af [mailto:af-boun...@afmug.com] On Behalf Of Eric Kuhnke Sent: May 19, 2016 3:28 PM To: af@afmug.com Subject: Re: [AFMUG] Data center temperatures Agreed from a end user perspective but not so much in a $$$/sq ft revenue perspective of a datacenter operator. With the very highest density/high power cabinets I've seen recently, the cable management is actually not so bad. For hypervisor platforms what used to be 4 x 1000BaseT connections and maybe a 5th cable for OOB in a previous generation design is now a few 10GbE and 40Gb links to a TOR switch over regular, thin, yellow spaghetti two strand singlemode. On Thu, May 19, 2016 at 12:23 PM, Josh Reynolds mailto:j...@kyneticwifi.com> > wrote: Extra wide cabinets are awesome for cable management. On Thu, May 19, 2016 at 12:42 PM, Paul Stewart mailto:p...@paulstewart.org> > wrote: > The cabinets are 50 or 52U in size – custom size I know for sure… extra wide > too which is nice > > > > When filled (pure SSD, almost 200TB raw capacity) they draw around 16kW of > power J > > > > > > From: Af [mailto:af-boun...@afmug.com <mailto:af-boun...@afmug.com> ] On > Behalf Of Eric Kuhnke > Sent: May 14, 2016 7:50 PM > To: af@afmug.com <mailto:af@afmug.com> > Subject: Re: [AFMUG] Data center temperatures > > > > How does a 44U cabinet need 208V 60A for storage arrays? > > In a 4U chassis the max hard drives (front and rear) is about 60 x 3.5"... > > Say each drive is 7.5W TDP, that's 450W of drives. Add another 200W for > controller/motherboard and fans. 650W in 4U. > > 44 / 4 = 11 > > Multply by 650 > > 7150W > > More realistically with a normal amount of drives (like 40 per 4U) a single > 208 30A is sufficient, > > 208 x 30 = 6240W > > Run at max 0.85 load on the circuit, so > > 6240 x 0.85 = 5304W > > In a really dense 2.5" environment all of the above is of course invalid, > you could probably need up to 7900W per cabinet > Then there's 52U cabinets as well... > > On May 13, 2016 6:16 PM, "Paul Stewart" <mailto:p...@paulstewart.org> > wrote: > > Yup … general trends on new data centers are pushing those temperatures > higher for efficiency but also with better designs .. > > > > One of our data centers runs at 78F and have no issues – each cabinet is > standard 208V 30A as you mention but can go per cabinet much higher if > needed (ie. 208V 60A for storage arrays) > > > > From: Af [mailto:af-boun...@afmug.com <mailto:af-boun...@afmug.com> ] On > Behalf Of Eric Kuhnke > Sent: May 11, 2016 5:15 PM > > > To: af@afmug.com <mailto:af@afmug.com> > Subject: Re: [AFMUG] Data center temperatures > > > > 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 <mailto:af...@kwisp.com> > 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 > > Sent: Wednesday, May 11, 2016 3:51 PM > > To: af@afmug.com <mailto:af@afmug.com> > > 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 <mailto:lesm...@ecsis.net> > wrote: > > On Wed May 11 2016 15:37, Josh Luthman wrote: >> Just cu
Re: [AFMUG] Data center temperatures
Extra wide 19” inch cabinets … they were specially ordered for “us” :) I can’t comment too much on the cooling as the design is rather custom in nature but can say that we’ve never had any issues there … From: Af [mailto:af-boun...@afmug.com] On Behalf Of Eric Kuhnke Sent: May 19, 2016 3:11 PM To: af@afmug.com Subject: Re: [AFMUG] Data center temperatures 52U and 23" rack, or just an extra-wide/roomy 19"? I'd like to see a photo of the hot/cold aisle set up for that, if they have a bunch of 208V 60A capable cabinets in a row On Thu, May 19, 2016 at 10:42 AM, Paul Stewart mailto:p...@paulstewart.org> > wrote: The cabinets are 50 or 52U in size – custom size I know for sure… extra wide too which is nice When filled (pure SSD, almost 200TB raw capacity) they draw around 16kW of power :) From: Af [mailto:af-boun...@afmug.com <mailto:af-boun...@afmug.com> ] On Behalf Of Eric Kuhnke Sent: May 14, 2016 7:50 PM To: af@afmug.com <mailto:af@afmug.com> Subject: Re: [AFMUG] Data center temperatures How does a 44U cabinet need 208V 60A for storage arrays? In a 4U chassis the max hard drives (front and rear) is about 60 x 3.5"... Say each drive is 7.5W TDP, that's 450W of drives. Add another 200W for controller/motherboard and fans. 650W in 4U. 44 / 4 = 11 Multply by 650 7150W More realistically with a normal amount of drives (like 40 per 4U) a single 208 30A is sufficient, 208 x 30 = 6240W Run at max 0.85 load on the circuit, so 6240 x 0.85 = 5304W In a really dense 2.5" environment all of the above is of course invalid, you could probably need up to 7900W per cabinet Then there's 52U cabinets as well... On May 13, 2016 6:16 PM, "Paul Stewart" mailto:p...@paulstewart.org> > wrote: Yup … general trends on new data centers are pushing those temperatures higher for efficiency but also with better designs .. One of our data centers runs at 78F and have no issues – each cabinet is standard 208V 30A as you mention but can go per cabinet much higher if needed (ie. 208V 60A for storage arrays) From: Af [mailto:af-boun...@afmug.com <mailto:af-boun...@afmug.com> ] On Behalf Of Eric Kuhnke Sent: May 11, 2016 5:15 PM To: af@afmug.com <mailto:af@afmug.com> Subject: Re: [AFMUG] Data center temperatures 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 mailto:af...@kwisp.com> > 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 <mailto:thatoneguyst...@gmail.com> Sent: Wednesday, May 11, 2016 3:51 PM To: af@afmug.com <mailto:af@afmug.com> 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 mailto:lesm...@ecsis.net> > 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 lesm...@ecsis.net <mailto:lesm...@ecsis.net> -- 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.
Re: [AFMUG] Data center temperatures
Varies. Used to be in the 60s?, but newer tech and newer designs can push that up near 100. - Mike Hammett Intelligent Computing Solutions Midwest Internet Exchange The Brothers WISP - Original Message - From: "Josh Luthman" To: af@afmug.com Sent: Wednesday, May 11, 2016 3:37:38 PM Subject: [AFMUG] Data center temperatures 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).
Re: [AFMUG] Data center temperatures
Agreed from a end user perspective but not so much in a $$$/sq ft revenue perspective of a datacenter operator. With the very highest density/high power cabinets I've seen recently, the cable management is actually not so bad. For hypervisor platforms what used to be 4 x 1000BaseT connections and maybe a 5th cable for OOB in a previous generation design is now a few 10GbE and 40Gb links to a TOR switch over regular, thin, yellow spaghetti two strand singlemode. On Thu, May 19, 2016 at 12:23 PM, Josh Reynolds wrote: > Extra wide cabinets are awesome for cable management. > > On Thu, May 19, 2016 at 12:42 PM, Paul Stewart > wrote: > > The cabinets are 50 or 52U in size – custom size I know for sure… extra > wide > > too which is nice > > > > > > > > When filled (pure SSD, almost 200TB raw capacity) they draw around 16kW > of > > power J > > > > > > > > > > > > From: Af [mailto:af-boun...@afmug.com] On Behalf Of Eric Kuhnke > > Sent: May 14, 2016 7:50 PM > > To: af@afmug.com > > Subject: Re: [AFMUG] Data center temperatures > > > > > > > > How does a 44U cabinet need 208V 60A for storage arrays? > > > > In a 4U chassis the max hard drives (front and rear) is about 60 x > 3.5"... > > > > Say each drive is 7.5W TDP, that's 450W of drives. Add another 200W for > > controller/motherboard and fans. 650W in 4U. > > > > 44 / 4 = 11 > > > > Multply by 650 > > > > 7150W > > > > More realistically with a normal amount of drives (like 40 per 4U) a > single > > 208 30A is sufficient, > > > > 208 x 30 = 6240W > > > > Run at max 0.85 load on the circuit, so > > > > 6240 x 0.85 = 5304W > > > > In a really dense 2.5" environment all of the above is of course invalid, > > you could probably need up to 7900W per cabinet > > Then there's 52U cabinets as well... > > > > On May 13, 2016 6:16 PM, "Paul Stewart" wrote: > > > > Yup … general trends on new data centers are pushing those temperatures > > higher for efficiency but also with better designs .. > > > > > > > > One of our data centers runs at 78F and have no issues – each cabinet is > > standard 208V 30A as you mention but can go per cabinet much higher if > > needed (ie. 208V 60A for storage arrays) > > > > > > > > From: Af [mailto:af-boun...@afmug.com] On Behalf Of Eric Kuhnke > > Sent: May 11, 2016 5:15 PM > > > > > > To: af@afmug.com > > Subject: Re: [AFMUG] Data center temperatures > > > > > > > > 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 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 > > > > Sent: Wednesday, May 11, 2016 3:51 PM > > > > To: af@afmug.com > > > > 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 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 > > lesm...@ecsis.net > > > > > > > > > > > > -- > > > > 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. > > > > >
Re: [AFMUG] Data center temperatures
Extra wide cabinets are awesome for cable management. On Thu, May 19, 2016 at 12:42 PM, Paul Stewart wrote: > The cabinets are 50 or 52U in size – custom size I know for sure… extra wide > too which is nice > > > > When filled (pure SSD, almost 200TB raw capacity) they draw around 16kW of > power J > > > > > > From: Af [mailto:af-boun...@afmug.com] On Behalf Of Eric Kuhnke > Sent: May 14, 2016 7:50 PM > To: af@afmug.com > Subject: Re: [AFMUG] Data center temperatures > > > > How does a 44U cabinet need 208V 60A for storage arrays? > > In a 4U chassis the max hard drives (front and rear) is about 60 x 3.5"... > > Say each drive is 7.5W TDP, that's 450W of drives. Add another 200W for > controller/motherboard and fans. 650W in 4U. > > 44 / 4 = 11 > > Multply by 650 > > 7150W > > More realistically with a normal amount of drives (like 40 per 4U) a single > 208 30A is sufficient, > > 208 x 30 = 6240W > > Run at max 0.85 load on the circuit, so > > 6240 x 0.85 = 5304W > > In a really dense 2.5" environment all of the above is of course invalid, > you could probably need up to 7900W per cabinet > Then there's 52U cabinets as well... > > On May 13, 2016 6:16 PM, "Paul Stewart" wrote: > > Yup … general trends on new data centers are pushing those temperatures > higher for efficiency but also with better designs .. > > > > One of our data centers runs at 78F and have no issues – each cabinet is > standard 208V 30A as you mention but can go per cabinet much higher if > needed (ie. 208V 60A for storage arrays) > > > > From: Af [mailto:af-boun...@afmug.com] On Behalf Of Eric Kuhnke > Sent: May 11, 2016 5:15 PM > > > To: af@afmug.com > Subject: Re: [AFMUG] Data center temperatures > > > > 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 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 > > Sent: Wednesday, May 11, 2016 3:51 PM > > To: af@afmug.com > > 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 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 > lesm...@ecsis.net > > > > > > -- > > 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. > >
Re: [AFMUG] Data center temperatures
52U and 23" rack, or just an extra-wide/roomy 19"? I'd like to see a photo of the hot/cold aisle set up for that, if they have a bunch of 208V 60A capable cabinets in a row On Thu, May 19, 2016 at 10:42 AM, Paul Stewart wrote: > The cabinets are 50 or 52U in size – custom size I know for sure… extra > wide too which is nice > > > > When filled (pure SSD, almost 200TB raw capacity) they draw around 16kW of > power J > > > > > > *From:* Af [mailto:af-boun...@afmug.com] *On Behalf Of *Eric Kuhnke > *Sent:* May 14, 2016 7:50 PM > *To:* af@afmug.com > *Subject:* Re: [AFMUG] Data center temperatures > > > > How does a 44U cabinet need 208V 60A for storage arrays? > > In a 4U chassis the max hard drives (front and rear) is about 60 x 3.5"... > > Say each drive is 7.5W TDP, that's 450W of drives. Add another 200W for > controller/motherboard and fans. 650W in 4U. > > 44 / 4 = 11 > > Multply by 650 > > 7150W > > More realistically with a normal amount of drives (like 40 per 4U) a > single 208 30A is sufficient, > > 208 x 30 = 6240W > > Run at max 0.85 load on the circuit, so > > 6240 x 0.85 = 5304W > > In a really dense 2.5" environment all of the above is of course invalid, > you could probably need up to 7900W per cabinet > Then there's 52U cabinets as well... > > On May 13, 2016 6:16 PM, "Paul Stewart" wrote: > > Yup … general trends on new data centers are pushing those temperatures > higher for efficiency but also with better designs .. > > > > One of our data centers runs at 78F and have no issues – each cabinet is > standard 208V 30A as you mention but can go per cabinet much higher if > needed (ie. 208V 60A for storage arrays) > > > > *From:* Af [mailto:af-boun...@afmug.com] *On Behalf Of *Eric Kuhnke > *Sent:* May 11, 2016 5:15 PM > > > *To:* af@afmug.com > *Subject:* Re: [AFMUG] Data center temperatures > > > > 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 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 > > *Sent:* Wednesday, May 11, 2016 3:51 PM > > *To:* af@afmug.com > > *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 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 > lesm...@ecsis.net > > > > > > -- > > 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. > > > >
Re: [AFMUG] Data center temperatures
Hahah… I’ve seen that several times especially in telco CO’s ;) From: Af [mailto:af-boun...@afmug.com] On Behalf Of Chuck McCown Sent: May 14, 2016 11:40 PM To: af@afmug.com Subject: Re: [AFMUG] Data center temperatures I remembering being at a data center on a hot summer day. Power went out, generator started. Things were fine... then all the air conditioners switched on at the same time. Actually stalled the generator. We had to put sequencers on the AC. From: Faisal Imtiaz <mailto:fai...@snappytelecom.net> Sent: Saturday, May 14, 2016 9:20 PM To: af@afmug.com <mailto:af@afmug.com> Subject: Re: [AFMUG] Data center temperatures FYI, Electrical Code (NECA) and most datacenters require the power not to be loaded beyond 80% of breaker capacity... i.e. 16amp draw on a 20amp circuit. Additionally, one also has to have head room on the power circuit to deal with start up draw (current rush). It's not pretty when you have a crap load of servers starting up all together :) Faisal Imtiaz Snappy Internet & Telecom 7266 SW 48 Street Miami, FL 33155 Tel: 305 663 5518 x 232 Help-desk: (305)663-5518 Option 2 or Email: supp...@snappytelecom.net <mailto:supp...@snappytelecom.net> _ From: "Eric Kuhnke" mailto:eric.kuh...@gmail.com> > To: af@afmug.com <mailto:af@afmug.com> Sent: Saturday, May 14, 2016 7:50:22 PM Subject: Re: [AFMUG] Data center temperatures How does a 44U cabinet need 208V 60A for storage arrays? In a 4U chassis the max hard drives (front and rear) is about 60 x 3.5"... Say each drive is 7.5W TDP, that's 450W of drives. Add another 200W for controller/motherboard and fans. 650W in 4U. 44 / 4 = 11 Multply by 650 7150W More realistically with a normal amount of drives (like 40 per 4U) a single 208 30A is sufficient, 208 x 30 = 6240W Run at max 0.85 load on the circuit, so 6240 x 0.85 = 5304W In a really dense 2.5" environment all of the above is of course invalid, you could probably need up to 7900W per cabinet Then there's 52U cabinets as well... On May 13, 2016 6:16 PM, "Paul Stewart" mailto:p...@paulstewart.org> > wrote: Yup … general trends on new data centers are pushing those temperatures higher for efficiency but also with better designs .. One of our data centers runs at 78F and have no issues – each cabinet is standard 208V 30A as you mention but can go per cabinet much higher if needed (ie. 208V 60A for storage arrays) From: Af [mailto:af-boun...@afmug.com <mailto:af-boun...@afmug.com> ] On Behalf Of Eric Kuhnke Sent: May 11, 2016 5:15 PM To: af@afmug.com <mailto:af@afmug.com> Subject: Re: [AFMUG] Data center temperatures 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 mailto:af...@kwisp.com> > 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 <mailto:thatoneguyst...@gmail.com> Sent: Wednesday, May 11, 2016 3:51 PM To: af@afmug.com <mailto:af@afmug.com> 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 mailto:lesm...@ecsis.net> > 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 lesm...@ecsis.net <mailto:lesm...@ecsis.net> -- 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.
Re: [AFMUG] Data center temperatures
The cabinets are 50 or 52U in size – custom size I know for sure… extra wide too which is nice When filled (pure SSD, almost 200TB raw capacity) they draw around 16kW of power :) From: Af [mailto:af-boun...@afmug.com] On Behalf Of Eric Kuhnke Sent: May 14, 2016 7:50 PM To: af@afmug.com Subject: Re: [AFMUG] Data center temperatures How does a 44U cabinet need 208V 60A for storage arrays? In a 4U chassis the max hard drives (front and rear) is about 60 x 3.5"... Say each drive is 7.5W TDP, that's 450W of drives. Add another 200W for controller/motherboard and fans. 650W in 4U. 44 / 4 = 11 Multply by 650 7150W More realistically with a normal amount of drives (like 40 per 4U) a single 208 30A is sufficient, 208 x 30 = 6240W Run at max 0.85 load on the circuit, so 6240 x 0.85 = 5304W In a really dense 2.5" environment all of the above is of course invalid, you could probably need up to 7900W per cabinet Then there's 52U cabinets as well... On May 13, 2016 6:16 PM, "Paul Stewart" mailto:p...@paulstewart.org> > wrote: Yup … general trends on new data centers are pushing those temperatures higher for efficiency but also with better designs .. One of our data centers runs at 78F and have no issues – each cabinet is standard 208V 30A as you mention but can go per cabinet much higher if needed (ie. 208V 60A for storage arrays) From: Af [mailto:af-boun...@afmug.com <mailto:af-boun...@afmug.com> ] On Behalf Of Eric Kuhnke Sent: May 11, 2016 5:15 PM To: af@afmug.com <mailto:af@afmug.com> Subject: Re: [AFMUG] Data center temperatures 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 mailto:af...@kwisp.com> > 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 <mailto:thatoneguyst...@gmail.com> Sent: Wednesday, May 11, 2016 3:51 PM To: af@afmug.com <mailto:af@afmug.com> 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 mailto:lesm...@ecsis.net> > 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 lesm...@ecsis.net <mailto:lesm...@ecsis.net> -- 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.
Re: [AFMUG] Data center temperatures
It happens during system start before raid array assembly and before OS boot On May 15, 2016 8:34 AM, "Josh Luthman" wrote: > How would staggered drive start up work? Or is the array not available > until the timer is done? > > Josh Luthman > Office: 937-552-2340 > Direct: 937-552-2343 > 1100 Wayne St > Suite 1337 > Troy, OH 45373 > On May 15, 2016 1:01 AM, "Josh Reynolds" wrote: > >> A lot of the dell servers I use, as well as a lot of the supermicro >> servers have that as well. Thankfully many of the RAID JBOD cards I >> use (softraid ftw, and zfs doesn't like it either) can also stagger >> drive startup. >> >> On Sat, May 14, 2016 at 11:13 PM, Faisal Imtiaz >> wrote: >> > It would be interesting to note that, we are putting in some new >> servers, >> > and in the bios these have a setting that delays a random amount of time >> > between 50 - 120seconds, before returning to power on state after a >> power >> > loss . >> > >> > :) >> > >> > Faisal Imtiaz >> > Snappy Internet & Telecom >> > 7266 SW 48 Street >> > Miami, FL 33155 >> > Tel: 305 663 5518 x 232 >> > >> > Help-desk: (305)663-5518 Option 2 or Email: supp...@snappytelecom.net >> > >> > >> > >> > From: "Chuck McCown" >> > To: af@afmug.com >> > Sent: Saturday, May 14, 2016 11:40:09 PM >> > >> > Subject: Re: [AFMUG] Data center temperatures >> > >> > I remembering being at a data center on a hot summer day. Power went >> out, >> > generator started. Things were fine... then all the air conditioners >> > switched on at the same time. Actually stalled the generator. We had >> to >> > put sequencers on the AC. >> > >> > From: Faisal Imtiaz >> > Sent: Saturday, May 14, 2016 9:20 PM >> > To: af@afmug.com >> > Subject: Re: [AFMUG] Data center temperatures >> > >> > FYI, Electrical Code (NECA) and most datacenters require the power not >> to be >> > loaded beyond 80% of breaker capacity... i.e. 16amp draw on a 20amp >> circuit. >> > >> > Additionally, one also has to have head room on the power circuit to >> deal >> > with start up draw (current rush). It's not pretty when you have a crap >> load >> > of servers starting up all together >> > >> > >> > :) >> > >> > Faisal Imtiaz >> > Snappy Internet & Telecom >> > 7266 SW 48 Street >> > Miami, FL 33155 >> > Tel: 305 663 5518 x 232 >> > >> > Help-desk: (305)663-5518 Option 2 or Email: supp...@snappytelecom.net >> > >> > >> > >> > From: "Eric Kuhnke" >> > To: af@afmug.com >> > Sent: Saturday, May 14, 2016 7:50:22 PM >> > Subject: Re: [AFMUG] Data center temperatures >> > >> > How does a 44U cabinet need 208V 60A for storage arrays? >> > >> > In a 4U chassis the max hard drives (front and rear) is about 60 x >> 3.5"... >> > >> > Say each drive is 7.5W TDP, that's 450W of drives. Add another 200W for >> > controller/motherboard and fans. 650W in 4U. >> > >> > 44 / 4 = 11 >> > >> > Multply by 650 >> > >> > 7150W >> > >> > More realistically with a normal amount of drives (like 40 per 4U) a >> single >> > 208 30A is sufficient, >> > >> > 208 x 30 = 6240W >> > >> > Run at max 0.85 load on the circuit, so >> > >> > 6240 x 0.85 = 5304W >> > >> > In a really dense 2.5" environment all of the above is of course >> invalid, >> > you could probably need up to 7900W per cabinet >> > Then there's 52U cabinets as well... >> > >> > On May 13, 2016 6:16 PM, "Paul Stewart" wrote: >> > >> > Yup … general trends on new data centers are pushing those temperatures >> > higher for efficiency but also with better designs .. >> > >> > >> > >> > One of our data centers runs at 78F and have no issues – each cabinet is >> > standard 208V 30A as you mention but can go per cabinet much higher if >> > needed (ie. 208V 60A for storage arrays) >> > >> > >> > >> > From: Af [mailto:af-boun...@afmug.com] On Behalf Of Eric Kuhnke >> > Sent: May 11, 2016 5:15
Re: [AFMUG] Data center temperatures
How would staggered drive start up work? Or is the array not available until the timer is done? Josh Luthman Office: 937-552-2340 Direct: 937-552-2343 1100 Wayne St Suite 1337 Troy, OH 45373 On May 15, 2016 1:01 AM, "Josh Reynolds" wrote: > A lot of the dell servers I use, as well as a lot of the supermicro > servers have that as well. Thankfully many of the RAID JBOD cards I > use (softraid ftw, and zfs doesn't like it either) can also stagger > drive startup. > > On Sat, May 14, 2016 at 11:13 PM, Faisal Imtiaz > wrote: > > It would be interesting to note that, we are putting in some new servers, > > and in the bios these have a setting that delays a random amount of time > > between 50 - 120seconds, before returning to power on state after a power > > loss . > > > > :) > > > > Faisal Imtiaz > > Snappy Internet & Telecom > > 7266 SW 48 Street > > Miami, FL 33155 > > Tel: 305 663 5518 x 232 > > > > Help-desk: (305)663-5518 Option 2 or Email: supp...@snappytelecom.net > > > > ____________ > > > > From: "Chuck McCown" > > To: af@afmug.com > > Sent: Saturday, May 14, 2016 11:40:09 PM > > > > Subject: Re: [AFMUG] Data center temperatures > > > > I remembering being at a data center on a hot summer day. Power went > out, > > generator started. Things were fine... then all the air conditioners > > switched on at the same time. Actually stalled the generator. We had > to > > put sequencers on the AC. > > > > From: Faisal Imtiaz > > Sent: Saturday, May 14, 2016 9:20 PM > > To: af@afmug.com > > Subject: Re: [AFMUG] Data center temperatures > > > > FYI, Electrical Code (NECA) and most datacenters require the power not > to be > > loaded beyond 80% of breaker capacity... i.e. 16amp draw on a 20amp > circuit. > > > > Additionally, one also has to have head room on the power circuit to deal > > with start up draw (current rush). It's not pretty when you have a crap > load > > of servers starting up all together > > > > > > :) > > > > Faisal Imtiaz > > Snappy Internet & Telecom > > 7266 SW 48 Street > > Miami, FL 33155 > > Tel: 305 663 5518 x 232 > > > > Help-desk: (305)663-5518 Option 2 or Email: supp...@snappytelecom.net > > > > > > > > From: "Eric Kuhnke" > > To: af@afmug.com > > Sent: Saturday, May 14, 2016 7:50:22 PM > > Subject: Re: [AFMUG] Data center temperatures > > > > How does a 44U cabinet need 208V 60A for storage arrays? > > > > In a 4U chassis the max hard drives (front and rear) is about 60 x > 3.5"... > > > > Say each drive is 7.5W TDP, that's 450W of drives. Add another 200W for > > controller/motherboard and fans. 650W in 4U. > > > > 44 / 4 = 11 > > > > Multply by 650 > > > > 7150W > > > > More realistically with a normal amount of drives (like 40 per 4U) a > single > > 208 30A is sufficient, > > > > 208 x 30 = 6240W > > > > Run at max 0.85 load on the circuit, so > > > > 6240 x 0.85 = 5304W > > > > In a really dense 2.5" environment all of the above is of course invalid, > > you could probably need up to 7900W per cabinet > > Then there's 52U cabinets as well... > > > > On May 13, 2016 6:16 PM, "Paul Stewart" wrote: > > > > Yup … general trends on new data centers are pushing those temperatures > > higher for efficiency but also with better designs .. > > > > > > > > One of our data centers runs at 78F and have no issues – each cabinet is > > standard 208V 30A as you mention but can go per cabinet much higher if > > needed (ie. 208V 60A for storage arrays) > > > > > > > > From: Af [mailto:af-boun...@afmug.com] On Behalf Of Eric Kuhnke > > Sent: May 11, 2016 5:15 PM > > > > > > To: af@afmug.com > > Subject: Re: [AFMUG] Data center temperatures > > > > > > > > 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
Re: [AFMUG] Data center temperatures
A lot of the dell servers I use, as well as a lot of the supermicro servers have that as well. Thankfully many of the RAID JBOD cards I use (softraid ftw, and zfs doesn't like it either) can also stagger drive startup. On Sat, May 14, 2016 at 11:13 PM, Faisal Imtiaz wrote: > It would be interesting to note that, we are putting in some new servers, > and in the bios these have a setting that delays a random amount of time > between 50 - 120seconds, before returning to power on state after a power > loss . > > :) > > Faisal Imtiaz > Snappy Internet & Telecom > 7266 SW 48 Street > Miami, FL 33155 > Tel: 305 663 5518 x 232 > > Help-desk: (305)663-5518 Option 2 or Email: supp...@snappytelecom.net > > > > From: "Chuck McCown" > To: af@afmug.com > Sent: Saturday, May 14, 2016 11:40:09 PM > > Subject: Re: [AFMUG] Data center temperatures > > I remembering being at a data center on a hot summer day. Power went out, > generator started. Things were fine... then all the air conditioners > switched on at the same time. Actually stalled the generator. We had to > put sequencers on the AC. > > From: Faisal Imtiaz > Sent: Saturday, May 14, 2016 9:20 PM > To: af@afmug.com > Subject: Re: [AFMUG] Data center temperatures > > FYI, Electrical Code (NECA) and most datacenters require the power not to be > loaded beyond 80% of breaker capacity... i.e. 16amp draw on a 20amp circuit. > > Additionally, one also has to have head room on the power circuit to deal > with start up draw (current rush). It's not pretty when you have a crap load > of servers starting up all together > > > :) > > Faisal Imtiaz > Snappy Internet & Telecom > 7266 SW 48 Street > Miami, FL 33155 > Tel: 305 663 5518 x 232 > > Help-desk: (305)663-5518 Option 2 or Email: supp...@snappytelecom.net > > > > From: "Eric Kuhnke" > To: af@afmug.com > Sent: Saturday, May 14, 2016 7:50:22 PM > Subject: Re: [AFMUG] Data center temperatures > > How does a 44U cabinet need 208V 60A for storage arrays? > > In a 4U chassis the max hard drives (front and rear) is about 60 x 3.5"... > > Say each drive is 7.5W TDP, that's 450W of drives. Add another 200W for > controller/motherboard and fans. 650W in 4U. > > 44 / 4 = 11 > > Multply by 650 > > 7150W > > More realistically with a normal amount of drives (like 40 per 4U) a single > 208 30A is sufficient, > > 208 x 30 = 6240W > > Run at max 0.85 load on the circuit, so > > 6240 x 0.85 = 5304W > > In a really dense 2.5" environment all of the above is of course invalid, > you could probably need up to 7900W per cabinet > Then there's 52U cabinets as well... > > On May 13, 2016 6:16 PM, "Paul Stewart" wrote: > > Yup … general trends on new data centers are pushing those temperatures > higher for efficiency but also with better designs .. > > > > One of our data centers runs at 78F and have no issues – each cabinet is > standard 208V 30A as you mention but can go per cabinet much higher if > needed (ie. 208V 60A for storage arrays) > > > > From: Af [mailto:af-boun...@afmug.com] On Behalf Of Eric Kuhnke > Sent: May 11, 2016 5:15 PM > > > To: af@afmug.com > Subject: Re: [AFMUG] Data center temperatures > > > > 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 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 > > Sent: Wednesday, May 11, 2016 3:51 PM > > To: af@afmug.com > > Subject: Re: [AFMUG] Data center temperatures > > > > apparently 72 is the the ideal for our
Re: [AFMUG] Data center temperatures
Neat :) Josh Luthman Office: 937-552-2340 Direct: 937-552-2343 1100 Wayne St Suite 1337 Troy, OH 45373 On May 15, 2016 12:26 AM, "Faisal Imtiaz" wrote: > nothing special, dell C2100's looks like these settings are getting to > be more common in stuff designed for high density data center install. > > Regards. > > Faisal Imtiaz > Snappy Internet & Telecom > 7266 SW 48 Street > Miami, FL 33155 > Tel: 305 663 5518 x 232 > > Help-desk: (305)663-5518 Option 2 or Email: supp...@snappytelecom.net > > -- > > *From: *"Josh Luthman" > *To: *af@afmug.com > *Sent: *Sunday, May 15, 2016 12:14:49 AM > *Subject: *Re: [AFMUG] Data center temperatures > > Wow that's cool! What kind of hardware are they? > > > Josh Luthman > Office: 937-552-2340 > Direct: 937-552-2343 > 1100 Wayne St > Suite 1337 > Troy, OH 45373 > > On Sun, May 15, 2016 at 12:13 AM, Faisal Imtiaz > wrote: > >> It would be interesting to note that, we are putting in some new servers, >> and in the bios these have a setting that delays a random amount of time >> between 50 - 120seconds, before returning to power on state after a power >> loss . >> >> :) >> >> Faisal Imtiaz >> Snappy Internet & Telecom >> 7266 SW 48 Street >> Miami, FL 33155 >> Tel: 305 663 5518 x 232 >> >> Help-desk: (305)663-5518 Option 2 or Email: supp...@snappytelecom.net >> >> -- >> >> *From: *"Chuck McCown" >> *To: *af@afmug.com >> *Sent: *Saturday, May 14, 2016 11:40:09 PM >> >> *Subject: *Re: [AFMUG] Data center temperatures >> >> I remembering being at a data center on a hot summer day. Power went >> out, generator started. Things were fine... then all the air conditioners >> switched on at the same time. Actually stalled the generator. We had to >> put sequencers on the AC. >> >> *From:* Faisal Imtiaz >> *Sent:* Saturday, May 14, 2016 9:20 PM >> *To:* af@afmug.com >> *Subject:* Re: [AFMUG] Data center temperatures >> >> FYI, Electrical Code (NECA) and most datacenters require the power not to >> be loaded beyond 80% of breaker capacity... i.e. 16amp draw on a 20amp >> circuit. >> >> Additionally, one also has to have head room on the power circuit to deal >> with start up draw (current rush). It's not pretty when you have a crap >> load of servers starting up all together >> >> >> :) >> >> Faisal Imtiaz >> Snappy Internet & Telecom >> 7266 SW 48 Street >> Miami, FL 33155 >> Tel: 305 663 5518 x 232 >> >> Help-desk: (305)663-5518 Option 2 or Email: supp...@snappytelecom.net >> >> -- >> >> *From: *"Eric Kuhnke" >> *To: *af@afmug.com >> *Sent: *Saturday, May 14, 2016 7:50:22 PM >> *Subject: *Re: [AFMUG] Data center temperatures >> >> How does a 44U cabinet need 208V 60A for storage arrays? >> >> In a 4U chassis the max hard drives (front and rear) is about 60 x 3.5"... >> >> Say each drive is 7.5W TDP, that's 450W of drives. Add another 200W for >> controller/motherboard and fans. 650W in 4U. >> >> 44 / 4 = 11 >> >> Multply by 650 >> >> 7150W >> >> More realistically with a normal amount of drives (like 40 per 4U) a >> single 208 30A is sufficient, >> >> 208 x 30 = 6240W >> >> Run at max 0.85 load on the circuit, so >> >> 6240 x 0.85 = 5304W >> >> In a really dense 2.5" environment all of the above is of course invalid, >> you could probably need up to 7900W per cabinet >> Then there's 52U cabinets as well... >> On May 13, 2016 6:16 PM, "Paul Stewart" wrote: >> >> Yup … general trends on new data centers are pushing those temperatures >> higher for efficiency but also with better designs .. >> >> >> >> One of our data centers runs at 78F and have no issues – each cabinet is >> standard 208V 30A as you mention but can go per cabinet much higher if >> needed (ie. 208V 60A for storage arrays) >> >> >> >> *From:* Af [mailto:af-boun...@afmug.com] *On Behalf Of *Eric Kuhnke >> *Sent:* May 11, 2016 5:15 PM >> >> *To:* af@afmug.com >> *Subject:* Re: [AFMUG] Data center temperatures >> >> >> >> 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 st
Re: [AFMUG] Data center temperatures
nothing special, dell C2100's looks like these settings are getting to be more common in stuff designed for high density data center install. Regards. Faisal Imtiaz Snappy Internet & Telecom 7266 SW 48 Street Miami, FL 33155 Tel: 305 663 5518 x 232 Help-desk: (305)663-5518 Option 2 or Email: supp...@snappytelecom.net > From: "Josh Luthman" > To: af@afmug.com > Sent: Sunday, May 15, 2016 12:14:49 AM > Subject: Re: [AFMUG] Data center temperatures > Wow that's cool! What kind of hardware are they? > Josh Luthman > Office: 937-552-2340 > Direct: 937-552-2343 > 1100 Wayne St > Suite 1337 > Troy, OH 45373 > On Sun, May 15, 2016 at 12:13 AM, Faisal Imtiaz < fai...@snappytelecom.net > > wrote: >> It would be interesting to note that, we are putting in some new servers, >> and in >> the bios these have a setting that delays a random amount of time between 50 >> - >> 120seconds, before returning to power on state after a power loss . >> :) >> Faisal Imtiaz >> Snappy Internet & Telecom >> 7266 SW 48 Street >> Miami, FL 33155 >> Tel: 305 663 5518 x 232 >> Help-desk: (305)663-5518 Option 2 or Email: supp...@snappytelecom.net >>> From: "Chuck McCown" < ch...@wbmfg.com > >>> To: af@afmug.com >>> Sent: Saturday, May 14, 2016 11:40:09 PM >>> Subject: Re: [AFMUG] Data center temperatures >>> I remembering being at a data center on a hot summer day. Power went out, >>> generator started. Things were fine... then all the air conditioners >>> switched >>> on at the same time. Actually stalled the generator. We had to put >>> sequencers >>> on the AC. >>> From: Faisal Imtiaz >>> Sent: Saturday, May 14, 2016 9:20 PM >>> To: af@afmug.com >>> Subject: Re: [AFMUG] Data center temperatures >>> FYI, Electrical Code (NECA) and most datacenters require the power not to be >>> loaded beyond 80% of breaker capacity... i.e. 16amp draw on a 20amp circuit. >>> Additionally, one also has to have head room on the power circuit to deal >>> with >>> start up draw (current rush). It's not pretty when you have a crap load of >>> servers starting up all together >>> :) >>> Faisal Imtiaz >>> Snappy Internet & Telecom >>> 7266 SW 48 Street >>> Miami, FL 33155 >>> Tel: 305 663 5518 x 232 >>> Help-desk: (305)663-5518 Option 2 or Email: supp...@snappytelecom.net >>>> From: "Eric Kuhnke" < eric.kuh...@gmail.com > >>>> To: af@afmug.com >>>> Sent: Saturday, May 14, 2016 7:50:22 PM >>>> Subject: Re: [AFMUG] Data center temperatures >>>> How does a 44U cabinet need 208V 60A for storage arrays? >>>> In a 4U chassis the max hard drives (front and rear) is about 60 x 3.5"... >>>> Say each drive is 7.5W TDP, that's 450W of drives. Add another 200W for >>>> controller/motherboard and fans. 650W in 4U. >>>> 44 / 4 = 11 >>>> Multply by 650 >>>> 7150W >>>> More realistically with a normal amount of drives (like 40 per 4U) a >>>> single 208 >>>> 30A is sufficient, >>>> 208 x 30 = 6240W >>>> Run at max 0.85 load on the circuit, so >>>> 6240 x 0.85 = 5304W >>>> In a really dense 2.5" environment all of the above is of course invalid, >>>> you >>>> could probably need up to 7900W per cabinet >>>> Then there's 52U cabinets as well... >>>> On May 13, 2016 6:16 PM, "Paul Stewart" < p...@paulstewart.org > wrote: >>>>> Yup … general trends on new data centers are pushing those temperatures >>>>> higher >>>>> for efficiency but also with better designs .. >>>>> One of our data centers runs at 78F and have no issues – each cabinet is >>>>> standard 208V 30A as you mention but can go per cabinet much higher if >>>>> needed >>>>> (ie. 208V 60A for storage arrays) >>>>> From: Af [mailto: af-boun...@afmug.com ] On Behalf Of Eric Kuhnke >>>>> Sent: May 11, 2016 5:15 PM >>>>> To: af@afmug.com >>>>> Subject: Re: [AFMUG] Data center temperatures >>>>> 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. &
Re: [AFMUG] Data center temperatures
Wow that's cool! What kind of hardware are they? Josh Luthman Office: 937-552-2340 Direct: 937-552-2343 1100 Wayne St Suite 1337 Troy, OH 45373 On Sun, May 15, 2016 at 12:13 AM, Faisal Imtiaz wrote: > It would be interesting to note that, we are putting in some new servers, > and in the bios these have a setting that delays a random amount of time > between 50 - 120seconds, before returning to power on state after a power > loss . > > :) > > Faisal Imtiaz > Snappy Internet & Telecom > 7266 SW 48 Street > Miami, FL 33155 > Tel: 305 663 5518 x 232 > > Help-desk: (305)663-5518 Option 2 or Email: supp...@snappytelecom.net > > -- > > *From: *"Chuck McCown" > *To: *af@afmug.com > *Sent: *Saturday, May 14, 2016 11:40:09 PM > > *Subject: *Re: [AFMUG] Data center temperatures > > I remembering being at a data center on a hot summer day. Power went > out, generator started. Things were fine... then all the air conditioners > switched on at the same time. Actually stalled the generator. We had to > put sequencers on the AC. > > *From:* Faisal Imtiaz > *Sent:* Saturday, May 14, 2016 9:20 PM > *To:* af@afmug.com > *Subject:* Re: [AFMUG] Data center temperatures > > FYI, Electrical Code (NECA) and most datacenters require the power not to > be loaded beyond 80% of breaker capacity... i.e. 16amp draw on a 20amp > circuit. > > Additionally, one also has to have head room on the power circuit to deal > with start up draw (current rush). It's not pretty when you have a crap > load of servers starting up all together > > > :) > > Faisal Imtiaz > Snappy Internet & Telecom > 7266 SW 48 Street > Miami, FL 33155 > Tel: 305 663 5518 x 232 > > Help-desk: (305)663-5518 Option 2 or Email: supp...@snappytelecom.net > > -- > > *From: *"Eric Kuhnke" > *To: *af@afmug.com > *Sent: *Saturday, May 14, 2016 7:50:22 PM > *Subject: *Re: [AFMUG] Data center temperatures > > How does a 44U cabinet need 208V 60A for storage arrays? > > In a 4U chassis the max hard drives (front and rear) is about 60 x 3.5"... > > Say each drive is 7.5W TDP, that's 450W of drives. Add another 200W for > controller/motherboard and fans. 650W in 4U. > > 44 / 4 = 11 > > Multply by 650 > > 7150W > > More realistically with a normal amount of drives (like 40 per 4U) a > single 208 30A is sufficient, > > 208 x 30 = 6240W > > Run at max 0.85 load on the circuit, so > > 6240 x 0.85 = 5304W > > In a really dense 2.5" environment all of the above is of course invalid, > you could probably need up to 7900W per cabinet > Then there's 52U cabinets as well... > On May 13, 2016 6:16 PM, "Paul Stewart" wrote: > > Yup … general trends on new data centers are pushing those temperatures > higher for efficiency but also with better designs .. > > > > One of our data centers runs at 78F and have no issues – each cabinet is > standard 208V 30A as you mention but can go per cabinet much higher if > needed (ie. 208V 60A for storage arrays) > > > > *From:* Af [mailto:af-boun...@afmug.com] *On Behalf Of *Eric Kuhnke > *Sent:* May 11, 2016 5:15 PM > > *To:* af@afmug.com > *Subject:* Re: [AFMUG] Data center temperatures > > > > 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 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 > > *Sent:* Wednesday, May 11, 2016 3:51 PM > > *To:* af@afmug.com > > *Subject:* Re: [AFMUG] Data center temperatures > > > > apparently 72 is the the ideal for our noc, i set our thermos
Re: [AFMUG] Data center temperatures
It would be interesting to note that, we are putting in some new servers, and in the bios these have a setting that delays a random amount of time between 50 - 120seconds, before returning to power on state after a power loss . :) Faisal Imtiaz Snappy Internet & Telecom 7266 SW 48 Street Miami, FL 33155 Tel: 305 663 5518 x 232 Help-desk: (305)663-5518 Option 2 or Email: supp...@snappytelecom.net > From: "Chuck McCown" > To: af@afmug.com > Sent: Saturday, May 14, 2016 11:40:09 PM > Subject: Re: [AFMUG] Data center temperatures > I remembering being at a data center on a hot summer day. Power went out, > generator started. Things were fine... then all the air conditioners switched > on at the same time. Actually stalled the generator. We had to put sequencers > on the AC. > From: Faisal Imtiaz > Sent: Saturday, May 14, 2016 9:20 PM > To: af@afmug.com > Subject: Re: [AFMUG] Data center temperatures > FYI, Electrical Code (NECA) and most datacenters require the power not to be > loaded beyond 80% of breaker capacity... i.e. 16amp draw on a 20amp circuit. > Additionally, one also has to have head room on the power circuit to deal with > start up draw (current rush). It's not pretty when you have a crap load of > servers starting up all together > :) > Faisal Imtiaz > Snappy Internet & Telecom > 7266 SW 48 Street > Miami, FL 33155 > Tel: 305 663 5518 x 232 > Help-desk: (305)663-5518 Option 2 or Email: supp...@snappytelecom.net >> From: "Eric Kuhnke" >> To: af@afmug.com >> Sent: Saturday, May 14, 2016 7:50:22 PM >> Subject: Re: [AFMUG] Data center temperatures >> How does a 44U cabinet need 208V 60A for storage arrays? >> In a 4U chassis the max hard drives (front and rear) is about 60 x 3.5"... >> Say each drive is 7.5W TDP, that's 450W of drives. Add another 200W for >> controller/motherboard and fans. 650W in 4U. >> 44 / 4 = 11 >> Multply by 650 >> 7150W >> More realistically with a normal amount of drives (like 40 per 4U) a single >> 208 >> 30A is sufficient, >> 208 x 30 = 6240W >> Run at max 0.85 load on the circuit, so >> 6240 x 0.85 = 5304W >> In a really dense 2.5" environment all of the above is of course invalid, you >> could probably need up to 7900W per cabinet >> Then there's 52U cabinets as well... >> On May 13, 2016 6:16 PM, "Paul Stewart" < p...@paulstewart.org > wrote: >>> Yup … general trends on new data centers are pushing those temperatures >>> higher >>> for efficiency but also with better designs .. >>> One of our data centers runs at 78F and have no issues – each cabinet is >>> standard 208V 30A as you mention but can go per cabinet much higher if >>> needed >>> (ie. 208V 60A for storage arrays) >>> From: Af [mailto: af-boun...@afmug.com ] On Behalf Of Eric Kuhnke >>> Sent: May 11, 2016 5:15 PM >>> To: af@afmug.com >>> Subject: Re: [AFMUG] Data center temperatures >>> 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 < af...@kwisp.com > 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 >>>> Sent: Wednesday, May 11, 2016 3:51 PM >>>> To: af@afmug.com >>>> 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 >>>> t
Re: [AFMUG] Data center temperatures
I remembering being at a data center on a hot summer day. Power went out, generator started. Things were fine... then all the air conditioners switched on at the same time. Actually stalled the generator. We had to put sequencers on the AC. From: Faisal Imtiaz Sent: Saturday, May 14, 2016 9:20 PM To: af@afmug.com Subject: Re: [AFMUG] Data center temperatures FYI, Electrical Code (NECA) and most datacenters require the power not to be loaded beyond 80% of breaker capacity... i.e. 16amp draw on a 20amp circuit. Additionally, one also has to have head room on the power circuit to deal with start up draw (current rush). It's not pretty when you have a crap load of servers starting up all together :) Faisal Imtiaz Snappy Internet & Telecom 7266 SW 48 Street Miami, FL 33155 Tel: 305 663 5518 x 232 Help-desk: (305)663-5518 Option 2 or Email: supp...@snappytelecom.net From: "Eric Kuhnke" To: af@afmug.com Sent: Saturday, May 14, 2016 7:50:22 PM Subject: Re: [AFMUG] Data center temperatures How does a 44U cabinet need 208V 60A for storage arrays? In a 4U chassis the max hard drives (front and rear) is about 60 x 3.5"... Say each drive is 7.5W TDP, that's 450W of drives. Add another 200W for controller/motherboard and fans. 650W in 4U. 44 / 4 = 11 Multply by 650 7150W More realistically with a normal amount of drives (like 40 per 4U) a single 208 30A is sufficient, 208 x 30 = 6240W Run at max 0.85 load on the circuit, so 6240 x 0.85 = 5304W In a really dense 2.5" environment all of the above is of course invalid, you could probably need up to 7900W per cabinet Then there's 52U cabinets as well... On May 13, 2016 6:16 PM, "Paul Stewart" wrote: Yup … general trends on new data centers are pushing those temperatures higher for efficiency but also with better designs .. One of our data centers runs at 78F and have no issues – each cabinet is standard 208V 30A as you mention but can go per cabinet much higher if needed (ie. 208V 60A for storage arrays) From: Af [mailto:af-boun...@afmug.com] On Behalf Of Eric Kuhnke Sent: May 11, 2016 5:15 PM To: af@afmug.com Subject: Re: [AFMUG] Data center temperatures 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 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 Sent: Wednesday, May 11, 2016 3:51 PM To: af@afmug.com 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 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 lesm...@ecsis.net -- 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.
Re: [AFMUG] Data center temperatures
FYI, Electrical Code (NECA) and most datacenters require the power not to be loaded beyond 80% of breaker capacity... i.e. 16amp draw on a 20amp circuit. Additionally, one also has to have head room on the power circuit to deal with start up draw (current rush). It's not pretty when you have a crap load of servers starting up all together :) Faisal Imtiaz Snappy Internet & Telecom 7266 SW 48 Street Miami, FL 33155 Tel: 305 663 5518 x 232 Help-desk: (305)663-5518 Option 2 or Email: supp...@snappytelecom.net > From: "Eric Kuhnke" > To: af@afmug.com > Sent: Saturday, May 14, 2016 7:50:22 PM > Subject: Re: [AFMUG] Data center temperatures > How does a 44U cabinet need 208V 60A for storage arrays? > In a 4U chassis the max hard drives (front and rear) is about 60 x 3.5"... > Say each drive is 7.5W TDP, that's 450W of drives. Add another 200W for > controller/motherboard and fans. 650W in 4U. > 44 / 4 = 11 > Multply by 650 > 7150W > More realistically with a normal amount of drives (like 40 per 4U) a single > 208 > 30A is sufficient, > 208 x 30 = 6240W > Run at max 0.85 load on the circuit, so > 6240 x 0.85 = 5304W > In a really dense 2.5" environment all of the above is of course invalid, you > could probably need up to 7900W per cabinet > Then there's 52U cabinets as well... > On May 13, 2016 6:16 PM, "Paul Stewart" < p...@paulstewart.org > wrote: >> Yup … general trends on new data centers are pushing those temperatures >> higher >> for efficiency but also with better designs .. >> One of our data centers runs at 78F and have no issues – each cabinet is >> standard 208V 30A as you mention but can go per cabinet much higher if needed >> (ie. 208V 60A for storage arrays) >> From: Af [mailto: af-boun...@afmug.com ] On Behalf Of Eric Kuhnke >> Sent: May 11, 2016 5:15 PM >> To: af@afmug.com >> Subject: Re: [AFMUG] Data center temperatures >> 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 < af...@kwisp.com > 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 >>> Sent: Wednesday, May 11, 2016 3:51 PM >>> To: af@afmug.com >>> 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 < lesm...@ecsis.net > 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 >>>> lesm...@ecsis.net >>> -- >>> 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.
Re: [AFMUG] Data center temperatures
Or if you had 3 or 4 mx960s per cabinet... 38A per power supply x 4 power supplies = 152A. 152A per chassis x 3 = 456A. An MX2020 is fun if running via DC power. ". A total of four PDMs can be installed into a router. Each DC PDM operates with up to nine separate feeds of either 60-amp or 80-amp current limit . The capacity of these feeds is relayed to system software through a switch located on the DC PDM." Oh, GPU clusters. Those are MASSIVE power hogs. I wonder how much power bitcoin clusters eat? These are all semi rare or rare cases of course :) On May 14, 2016 6:59 PM, "Seth Mattinen" wrote: > On 5/14/16 16:50, Eric Kuhnke wrote: > >> In a really dense 2.5" environment all of the above is of course >> invalid, you could probably need up to 7900W per cabinet >> > > > I have customers that peak at 10kW per cabinet, but that's HPC, not > storage. > > ~Seth >
Re: [AFMUG] Data center temperatures
Unless you're Dropbox, then you have all kinds of drives crammed in custom enclosures. "Basically Diskotech stores 1PB in 18" × 6" × 42" = 4,536 cubic inch volume, which is 10% bigger than standard 7U. [Backblaze] is [storing] 180TB in 4U. ... Doing the math reveals that Dropbox is basically packing 793TB in 4U. … Diskotech is about 30% bigger in volume than [Backblaze] Storage Pod 5.0 but with 470% more storage." On May 14, 2016 6:50 PM, "Eric Kuhnke" wrote: > How does a 44U cabinet need 208V 60A for storage arrays? > > In a 4U chassis the max hard drives (front and rear) is about 60 x 3.5"... > > Say each drive is 7.5W TDP, that's 450W of drives. Add another 200W for > controller/motherboard and fans. 650W in 4U. > > 44 / 4 = 11 > > Multply by 650 > > 7150W > > More realistically with a normal amount of drives (like 40 per 4U) a > single 208 30A is sufficient, > > 208 x 30 = 6240W > > Run at max 0.85 load on the circuit, so > > 6240 x 0.85 = 5304W > > In a really dense 2.5" environment all of the above is of course invalid, > you could probably need up to 7900W per cabinet > Then there's 52U cabinets as well... > On May 13, 2016 6:16 PM, "Paul Stewart" wrote: > > Yup … general trends on new data centers are pushing those temperatures > higher for efficiency but also with better designs .. > > > > One of our data centers runs at 78F and have no issues – each cabinet is > standard 208V 30A as you mention but can go per cabinet much higher if > needed (ie. 208V 60A for storage arrays) > > > > *From:* Af [mailto:af-boun...@afmug.com] *On Behalf Of *Eric Kuhnke > *Sent:* May 11, 2016 5:15 PM > > *To:* af@afmug.com > *Subject:* Re: [AFMUG] Data center temperatures > > > > 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 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 > > *Sent:* Wednesday, May 11, 2016 3:51 PM > > *To:* af@afmug.com > > *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 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 > lesm...@ecsis.net > > > > > > -- > > 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. > > > >
Re: [AFMUG] Data center temperatures
On 5/14/16 16:50, Eric Kuhnke wrote: In a really dense 2.5" environment all of the above is of course invalid, you could probably need up to 7900W per cabinet I have customers that peak at 10kW per cabinet, but that's HPC, not storage. ~Seth
Re: [AFMUG] Data center temperatures
How does a 44U cabinet need 208V 60A for storage arrays? In a 4U chassis the max hard drives (front and rear) is about 60 x 3.5"... Say each drive is 7.5W TDP, that's 450W of drives. Add another 200W for controller/motherboard and fans. 650W in 4U. 44 / 4 = 11 Multply by 650 7150W More realistically with a normal amount of drives (like 40 per 4U) a single 208 30A is sufficient, 208 x 30 = 6240W Run at max 0.85 load on the circuit, so 6240 x 0.85 = 5304W In a really dense 2.5" environment all of the above is of course invalid, you could probably need up to 7900W per cabinet Then there's 52U cabinets as well... On May 13, 2016 6:16 PM, "Paul Stewart" wrote: Yup … general trends on new data centers are pushing those temperatures higher for efficiency but also with better designs .. One of our data centers runs at 78F and have no issues – each cabinet is standard 208V 30A as you mention but can go per cabinet much higher if needed (ie. 208V 60A for storage arrays) *From:* Af [mailto:af-boun...@afmug.com] *On Behalf Of *Eric Kuhnke *Sent:* May 11, 2016 5:15 PM *To:* af@afmug.com *Subject:* Re: [AFMUG] Data center temperatures 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 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 *Sent:* Wednesday, May 11, 2016 3:51 PM *To:* af@afmug.com *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 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 lesm...@ecsis.net -- 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.
Re: [AFMUG] Data center temperatures
Yup … general trends on new data centers are pushing those temperatures higher for efficiency but also with better designs .. One of our data centers runs at 78F and have no issues – each cabinet is standard 208V 30A as you mention but can go per cabinet much higher if needed (ie. 208V 60A for storage arrays) From: Af [mailto:af-boun...@afmug.com] On Behalf Of Eric Kuhnke Sent: May 11, 2016 5:15 PM To: af@afmug.com Subject: Re: [AFMUG] Data center temperatures 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 mailto:af...@kwisp.com> > 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 <mailto:thatoneguyst...@gmail.com> Sent: Wednesday, May 11, 2016 3:51 PM To: af@afmug.com <mailto:af@afmug.com> 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 mailto:lesm...@ecsis.net> > 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 lesm...@ecsis.net <mailto:lesm...@ecsis.net> -- 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.
Re: [AFMUG] Data center temperatures
We settled ours at 72 degrees many years ago, after trying different ranges for a while. We have a 2nd “standby” AC that we can kick in if we get notices of server room temperature. We get notified by a SiteMonitor and by the Honeywell thermostat also of significant changes in temperature From: Af [mailto:af-boun...@afmug.com] On Behalf Of That One Guy /sarcasm Sent: Thursday, May 12, 2016 12:31 AM To: af@afmug.com Subject: Re: [AFMUG] Data center temperatures We did a 4.9ghz project for a municipality once, their server room was like a freezer, you could see your breath and everything On May 11, 2016 10:43 PM, "Travis Johnson" mailto:t...@ida.net>> wrote: We always kept our NOC temps around 72-74F... mainly because that would give us time if an A/C unit failed (or switched off due to power failure, etc.) to get physically to the NOC before temps reached above 100F (which did happen a few times in my 16 years). Servers start shutting down when the air intake hits about 105F. LOL Travis On 5/11/2016 5:53 PM, Robert Andrews wrote: Exactly... Hence our love for the old MAE East... On 05/11/2016 04:47 PM, Josh Luthman wrote: Parking garages are generally hotter then hell or balls cold in my experience. Josh Luthman Office: 937-552-2340 Direct: 937-552-2343 1100 Wayne St Suite 1337 Troy, OH 45373 On Wed, May 11, 2016 at 7:31 PM, Eric Kuhnke mailto:eric.kuh...@gmail.com> <mailto:eric.kuh...@gmail.com<mailto:eric.kuh...@gmail.com>>> wrote: The temperature sensor location on a 6503/6506/6509 isn't really at the 'raw' air intake, so it's showing warmer than it should be, but yes that cabinet gets warm... It's a couple of hundred watts heat load in a ventilated box. I would estimate the actual intake air temperature if you were to measure it manually with a thermometer is 26-27C on the right side of the 6503 as you're facing the front. The parking garage is pretty much the ambient air temperature of the city it's located in, but not exposed directly to sunlight. On Wed, May 11, 2016 at 4:26 PM, Josh Luthman mailto:j...@imaginenetworksllc.com> <mailto:j...@imaginenetworksllc.com<mailto:j...@imaginenetworksllc.com>>> wrote: 104F air intake? No way!!! On May 11, 2016 7:15 PM, "Eric Kuhnke" mailto:eric.kuh...@gmail.com> <mailto:eric.kuh...@gmail.com<mailto:eric.kuh...@gmail.com>>> wrote: Here's a chart from 2014, it's the air intake temperature sensor for a cisco 6503 in a wall mounted cabinet 9' in the air in a parking garage. The daily cycles are the ambient air temperature in the garage changing. On Wed, May 11, 2016 at 4:04 PM, Keefe John mailto:keefe...@ethoplex.com> <mailto:keefe...@ethoplex.com<mailto:keefe...@ethoplex.com>>> wrote: We do 75 degrees On 5/11/2016 5:51 PM, Robert Andrews wrote: This is related to the lubricant that is used in the drives. Seagate is to blame.. They discovered higher spindle speeds require lubricants that like higher temps... There is a secondary effect due to the way that magnetized materials flip and hold at higher temps. Again, my data may be old as I worked in that industry 20 years ago.. On 05/11/2016 02:58 PM, Chuck McCown wrote: Yep, hot is good according to Google. Somewhere there is a rotating media study that shows they last longer at higher temps. Who woulda thunk. -Original Message- From: Josh Reynolds Sent: Wednesday, May 11, 2016 2:48 PM To: af@afmug.com<mailto:af@afmug.com> <mailto:af@afmug.com<mailto:af@afmug.com>> Subject: Re: [AFMUG] Data center temperatures Ours is at 68deg F, and we monitor dewpoint and humidity ranges. However... http://www.geek.com/chips/googles-most-efficient-data-center-runs-at-95-degrees-1478473/ On Wed, May 11, 2016 at 3:37 PM, Josh Luthman mailto:j...@imaginenetworksllc.com> <mailto:j...@imaginenetworksllc.com<mailto:j...@imaginenetworksllc.com>>> 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).
Re: [AFMUG] Data center temperatures
We did a 4.9ghz project for a municipality once, their server room was like a freezer, you could see your breath and everything On May 11, 2016 10:43 PM, "Travis Johnson" wrote: > We always kept our NOC temps around 72-74F... mainly because that would > give us time if an A/C unit failed (or switched off due to power failure, > etc.) to get physically to the NOC before temps reached above 100F (which > did happen a few times in my 16 years). Servers start shutting down when > the air intake hits about 105F. LOL > > Travis > > > On 5/11/2016 5:53 PM, Robert Andrews wrote: > >> Exactly... Hence our love for the old MAE East... >> >> On 05/11/2016 04:47 PM, Josh Luthman wrote: >> >>> Parking garages are generally hotter then hell or balls cold in my >>> experience. >>> >>> >>> Josh Luthman >>> Office: 937-552-2340 >>> Direct: 937-552-2343 >>> 1100 Wayne St >>> Suite 1337 >>> Troy, OH 45373 >>> >>> On Wed, May 11, 2016 at 7:31 PM, Eric Kuhnke >> <mailto:eric.kuh...@gmail.com>> wrote: >>> >>> The temperature sensor location on a 6503/6506/6509 isn't really at >>> the 'raw' air intake, so it's showing warmer than it should be, but >>> yes that cabinet gets warm... It's a couple of hundred watts heat >>> load in a ventilated box. I would estimate the actual intake air >>> temperature if you were to measure it manually with a thermometer is >>> 26-27C on the right side of the 6503 as you're facing the front. >>> >>> The parking garage is pretty much the ambient air temperature of the >>> city it's located in, but not exposed directly to sunlight. >>> >>> On Wed, May 11, 2016 at 4:26 PM, Josh Luthman >>> mailto:j...@imaginenetworksllc.com>> >>> wrote: >>> >>> 104F air intake? No way!!! >>> >>> On May 11, 2016 7:15 PM, "Eric Kuhnke" >> <mailto:eric.kuh...@gmail.com>> wrote: >>> >>> Here's a chart from 2014, it's the air intake temperature >>> sensor for a cisco 6503 in a wall mounted cabinet 9' in the >>> air in a parking garage. The daily cycles are the ambient >>> air temperature in the garage changing. >>> >>> >>> >>> On Wed, May 11, 2016 at 4:04 PM, Keefe John >>> mailto:keefe...@ethoplex.com>> >>> wrote: >>> >>> We do 75 degrees >>> >>> >>> >>> On 5/11/2016 5:51 PM, Robert Andrews wrote: >>> >>> This is related to the lubricant that is used in the >>> drives. Seagate is to blame.. They discovered >>> higher spindle speeds require lubricants that like >>> higher temps... There is a secondary effect due to >>> the way that magnetized materials flip and hold at >>> higher temps. Again, my data may be old as I >>> worked in that industry 20 years ago.. >>> >>> On 05/11/2016 02:58 PM, Chuck McCown wrote: >>> >>> Yep, hot is good according to Google. Somewhere >>> there is a rotating >>> media study that shows they last longer at >>> higher temps. Who woulda thunk. >>> >>> -Original Message- From: Josh Reynolds >>> Sent: Wednesday, May 11, 2016 2:48 PM >>> To: af@afmug.com <mailto:af@afmug.com> >>> Subject: Re: [AFMUG] Data center temperatures >>> >>> Ours is at 68deg F, and we monitor dewpoint and >>> humidity ranges. >>> >>> However... >>> >>> http://www.geek.com/chips/googles-most-efficient-data-center-runs-at-95-degrees-1478473/ >>> >>> >>> >>> On Wed, May 11, 2016 at 3:37 PM, Josh Luthman >>> >> <mailto:j...@imaginenetworksllc.com>> 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). >>> >>> >>> >>> >>> >>> >>> >> >
Re: [AFMUG] Data center temperatures
We always kept our NOC temps around 72-74F... mainly because that would give us time if an A/C unit failed (or switched off due to power failure, etc.) to get physically to the NOC before temps reached above 100F (which did happen a few times in my 16 years). Servers start shutting down when the air intake hits about 105F. LOL Travis On 5/11/2016 5:53 PM, Robert Andrews wrote: Exactly... Hence our love for the old MAE East... On 05/11/2016 04:47 PM, Josh Luthman wrote: Parking garages are generally hotter then hell or balls cold in my experience. Josh Luthman Office: 937-552-2340 Direct: 937-552-2343 1100 Wayne St Suite 1337 Troy, OH 45373 On Wed, May 11, 2016 at 7:31 PM, Eric Kuhnke mailto:eric.kuh...@gmail.com>> wrote: The temperature sensor location on a 6503/6506/6509 isn't really at the 'raw' air intake, so it's showing warmer than it should be, but yes that cabinet gets warm... It's a couple of hundred watts heat load in a ventilated box. I would estimate the actual intake air temperature if you were to measure it manually with a thermometer is 26-27C on the right side of the 6503 as you're facing the front. The parking garage is pretty much the ambient air temperature of the city it's located in, but not exposed directly to sunlight. On Wed, May 11, 2016 at 4:26 PM, Josh Luthman mailto:j...@imaginenetworksllc.com>> wrote: 104F air intake? No way!!! On May 11, 2016 7:15 PM, "Eric Kuhnke" mailto:eric.kuh...@gmail.com>> wrote: Here's a chart from 2014, it's the air intake temperature sensor for a cisco 6503 in a wall mounted cabinet 9' in the air in a parking garage. The daily cycles are the ambient air temperature in the garage changing. On Wed, May 11, 2016 at 4:04 PM, Keefe John mailto:keefe...@ethoplex.com>> wrote: We do 75 degrees On 5/11/2016 5:51 PM, Robert Andrews wrote: This is related to the lubricant that is used in the drives. Seagate is to blame.. They discovered higher spindle speeds require lubricants that like higher temps... There is a secondary effect due to the way that magnetized materials flip and hold at higher temps. Again, my data may be old as I worked in that industry 20 years ago.. On 05/11/2016 02:58 PM, Chuck McCown wrote: Yep, hot is good according to Google. Somewhere there is a rotating media study that shows they last longer at higher temps. Who woulda thunk. -Original Message- From: Josh Reynolds Sent: Wednesday, May 11, 2016 2:48 PM To: af@afmug.com <mailto:af@afmug.com> Subject: Re: [AFMUG] Data center temperatures Ours is at 68deg F, and we monitor dewpoint and humidity ranges. However... http://www.geek.com/chips/googles-most-efficient-data-center-runs-at-95-degrees-1478473/ On Wed, May 11, 2016 at 3:37 PM, Josh Luthman mailto:j...@imaginenetworksllc.com>> 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).
Re: [AFMUG] Data center temperatures
Exactly... Hence our love for the old MAE East... On 05/11/2016 04:47 PM, Josh Luthman wrote: Parking garages are generally hotter then hell or balls cold in my experience. Josh Luthman Office: 937-552-2340 Direct: 937-552-2343 1100 Wayne St Suite 1337 Troy, OH 45373 On Wed, May 11, 2016 at 7:31 PM, Eric Kuhnke mailto:eric.kuh...@gmail.com>> wrote: The temperature sensor location on a 6503/6506/6509 isn't really at the 'raw' air intake, so it's showing warmer than it should be, but yes that cabinet gets warm... It's a couple of hundred watts heat load in a ventilated box. I would estimate the actual intake air temperature if you were to measure it manually with a thermometer is 26-27C on the right side of the 6503 as you're facing the front. The parking garage is pretty much the ambient air temperature of the city it's located in, but not exposed directly to sunlight. On Wed, May 11, 2016 at 4:26 PM, Josh Luthman mailto:j...@imaginenetworksllc.com>> wrote: 104F air intake? No way!!! On May 11, 2016 7:15 PM, "Eric Kuhnke" mailto:eric.kuh...@gmail.com>> wrote: Here's a chart from 2014, it's the air intake temperature sensor for a cisco 6503 in a wall mounted cabinet 9' in the air in a parking garage. The daily cycles are the ambient air temperature in the garage changing. On Wed, May 11, 2016 at 4:04 PM, Keefe John mailto:keefe...@ethoplex.com>> wrote: We do 75 degrees On 5/11/2016 5:51 PM, Robert Andrews wrote: This is related to the lubricant that is used in the drives. Seagate is to blame.. They discovered higher spindle speeds require lubricants that like higher temps... There is a secondary effect due to the way that magnetized materials flip and hold at higher temps. Again, my data may be old as I worked in that industry 20 years ago.. On 05/11/2016 02:58 PM, Chuck McCown wrote: Yep, hot is good according to Google. Somewhere there is a rotating media study that shows they last longer at higher temps. Who woulda thunk. -Original Message- From: Josh Reynolds Sent: Wednesday, May 11, 2016 2:48 PM To: af@afmug.com <mailto:af@afmug.com> Subject: Re: [AFMUG] Data center temperatures Ours is at 68deg F, and we monitor dewpoint and humidity ranges. However... http://www.geek.com/chips/googles-most-efficient-data-center-runs-at-95-degrees-1478473/ On Wed, May 11, 2016 at 3:37 PM, Josh Luthman mailto:j...@imaginenetworksllc.com>> 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).
Re: [AFMUG] Data center temperatures
Parking garages are generally hotter then hell or balls cold in my experience. Josh Luthman Office: 937-552-2340 Direct: 937-552-2343 1100 Wayne St Suite 1337 Troy, OH 45373 On Wed, May 11, 2016 at 7:31 PM, Eric Kuhnke wrote: > The temperature sensor location on a 6503/6506/6509 isn't really at the > 'raw' air intake, so it's showing warmer than it should be, but yes that > cabinet gets warm... It's a couple of hundred watts heat load in a > ventilated box. I would estimate the actual intake air temperature if you > were to measure it manually with a thermometer is 26-27C on the right side > of the 6503 as you're facing the front. > > The parking garage is pretty much the ambient air temperature of the city > it's located in, but not exposed directly to sunlight. > > On Wed, May 11, 2016 at 4:26 PM, Josh Luthman > wrote: > >> 104F air intake? No way!!! >> On May 11, 2016 7:15 PM, "Eric Kuhnke" wrote: >> >>> Here's a chart from 2014, it's the air intake temperature sensor for a >>> cisco 6503 in a wall mounted cabinet 9' in the air in a parking garage. The >>> daily cycles are the ambient air temperature in the garage changing. >>> >>> >>> >>> On Wed, May 11, 2016 at 4:04 PM, Keefe John >>> wrote: >>> >>>> We do 75 degrees >>>> >>>> >>>> >>>> On 5/11/2016 5:51 PM, Robert Andrews wrote: >>>> >>>>> This is related to the lubricant that is used in the drives. Seagate >>>>> is to blame.. They discovered higher spindle speeds require lubricants >>>>> that like higher temps... There is a secondary effect due to the way >>>>> that >>>>> magnetized materials flip and hold at higher temps. Again, my data may >>>>> be >>>>> old as I worked in that industry 20 years ago.. >>>>> >>>>> On 05/11/2016 02:58 PM, Chuck McCown wrote: >>>>> >>>>>> Yep, hot is good according to Google. Somewhere there is a rotating >>>>>> media study that shows they last longer at higher temps. Who woulda >>>>>> thunk. >>>>>> >>>>>> -Original Message- From: Josh Reynolds >>>>>> Sent: Wednesday, May 11, 2016 2:48 PM >>>>>> To: af@afmug.com >>>>>> Subject: Re: [AFMUG] Data center temperatures >>>>>> >>>>>> Ours is at 68deg F, and we monitor dewpoint and humidity ranges. >>>>>> >>>>>> However... >>>>>> >>>>>> http://www.geek.com/chips/googles-most-efficient-data-center-runs-at-95-degrees-1478473/ >>>>>> >>>>>> >>>>>> On Wed, May 11, 2016 at 3:37 PM, 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). >>>>>>> >>>>>> >>>>>> >>>> >>> >
Re: [AFMUG] Data center temperatures
The temperature sensor location on a 6503/6506/6509 isn't really at the 'raw' air intake, so it's showing warmer than it should be, but yes that cabinet gets warm... It's a couple of hundred watts heat load in a ventilated box. I would estimate the actual intake air temperature if you were to measure it manually with a thermometer is 26-27C on the right side of the 6503 as you're facing the front. The parking garage is pretty much the ambient air temperature of the city it's located in, but not exposed directly to sunlight. On Wed, May 11, 2016 at 4:26 PM, Josh Luthman wrote: > 104F air intake? No way!!! > On May 11, 2016 7:15 PM, "Eric Kuhnke" wrote: > >> Here's a chart from 2014, it's the air intake temperature sensor for a >> cisco 6503 in a wall mounted cabinet 9' in the air in a parking garage. The >> daily cycles are the ambient air temperature in the garage changing. >> >> >> >> On Wed, May 11, 2016 at 4:04 PM, Keefe John >> wrote: >> >>> We do 75 degrees >>> >>> >>> >>> On 5/11/2016 5:51 PM, Robert Andrews wrote: >>> >>>> This is related to the lubricant that is used in the drives. Seagate >>>> is to blame.. They discovered higher spindle speeds require lubricants >>>> that like higher temps... There is a secondary effect due to the way that >>>> magnetized materials flip and hold at higher temps. Again, my data may be >>>> old as I worked in that industry 20 years ago.. >>>> >>>> On 05/11/2016 02:58 PM, Chuck McCown wrote: >>>> >>>>> Yep, hot is good according to Google. Somewhere there is a rotating >>>>> media study that shows they last longer at higher temps. Who woulda >>>>> thunk. >>>>> >>>>> -Original Message- From: Josh Reynolds >>>>> Sent: Wednesday, May 11, 2016 2:48 PM >>>>> To: af@afmug.com >>>>> Subject: Re: [AFMUG] Data center temperatures >>>>> >>>>> Ours is at 68deg F, and we monitor dewpoint and humidity ranges. >>>>> >>>>> However... >>>>> >>>>> http://www.geek.com/chips/googles-most-efficient-data-center-runs-at-95-degrees-1478473/ >>>>> >>>>> >>>>> On Wed, May 11, 2016 at 3:37 PM, 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). >>>>>> >>>>> >>>>> >>> >>
Re: [AFMUG] Data center temperatures
104F air intake? No way!!! On May 11, 2016 7:15 PM, "Eric Kuhnke" wrote: > Here's a chart from 2014, it's the air intake temperature sensor for a > cisco 6503 in a wall mounted cabinet 9' in the air in a parking garage. The > daily cycles are the ambient air temperature in the garage changing. > > > > On Wed, May 11, 2016 at 4:04 PM, Keefe John wrote: > >> We do 75 degrees >> >> >> >> On 5/11/2016 5:51 PM, Robert Andrews wrote: >> >>> This is related to the lubricant that is used in the drives. Seagate >>> is to blame.. They discovered higher spindle speeds require lubricants >>> that like higher temps... There is a secondary effect due to the way that >>> magnetized materials flip and hold at higher temps. Again, my data may be >>> old as I worked in that industry 20 years ago.. >>> >>> On 05/11/2016 02:58 PM, Chuck McCown wrote: >>> >>>> Yep, hot is good according to Google. Somewhere there is a rotating >>>> media study that shows they last longer at higher temps. Who woulda >>>> thunk. >>>> >>>> -Original Message- From: Josh Reynolds >>>> Sent: Wednesday, May 11, 2016 2:48 PM >>>> To: af@afmug.com >>>> Subject: Re: [AFMUG] Data center temperatures >>>> >>>> Ours is at 68deg F, and we monitor dewpoint and humidity ranges. >>>> >>>> However... >>>> >>>> http://www.geek.com/chips/googles-most-efficient-data-center-runs-at-95-degrees-1478473/ >>>> >>>> >>>> On Wed, May 11, 2016 at 3:37 PM, 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). >>>>> >>>> >>>> >> >
Re: [AFMUG] Data center temperatures
We do 75 degrees On 5/11/2016 5:51 PM, Robert Andrews wrote: This is related to the lubricant that is used in the drives. Seagate is to blame.. They discovered higher spindle speeds require lubricants that like higher temps... There is a secondary effect due to the way that magnetized materials flip and hold at higher temps. Again, my data may be old as I worked in that industry 20 years ago.. On 05/11/2016 02:58 PM, Chuck McCown wrote: Yep, hot is good according to Google. Somewhere there is a rotating media study that shows they last longer at higher temps. Who woulda thunk. -Original Message- From: Josh Reynolds Sent: Wednesday, May 11, 2016 2:48 PM To: af@afmug.com Subject: Re: [AFMUG] Data center temperatures Ours is at 68deg F, and we monitor dewpoint and humidity ranges. However... http://www.geek.com/chips/googles-most-efficient-data-center-runs-at-95-degrees-1478473/ On Wed, May 11, 2016 at 3:37 PM, 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).
Re: [AFMUG] Data center temperatures
This is related to the lubricant that is used in the drives. Seagate is to blame.. They discovered higher spindle speeds require lubricants that like higher temps... There is a secondary effect due to the way that magnetized materials flip and hold at higher temps. Again, my data may be old as I worked in that industry 20 years ago.. On 05/11/2016 02:58 PM, Chuck McCown wrote: Yep, hot is good according to Google. Somewhere there is a rotating media study that shows they last longer at higher temps. Who woulda thunk. -Original Message- From: Josh Reynolds Sent: Wednesday, May 11, 2016 2:48 PM To: af@afmug.com Subject: Re: [AFMUG] Data center temperatures Ours is at 68deg F, and we monitor dewpoint and humidity ranges. However... http://www.geek.com/chips/googles-most-efficient-data-center-runs-at-95-degrees-1478473/ On Wed, May 11, 2016 at 3:37 PM, 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).
Re: [AFMUG] Data center temperatures
Yep, hot is good according to Google. Somewhere there is a rotating media study that shows they last longer at higher temps. Who woulda thunk. -Original Message- From: Josh Reynolds Sent: Wednesday, May 11, 2016 2:48 PM To: af@afmug.com Subject: Re: [AFMUG] Data center temperatures Ours is at 68deg F, and we monitor dewpoint and humidity ranges. However... http://www.geek.com/chips/googles-most-efficient-data-center-runs-at-95-degrees-1478473/ On Wed, May 11, 2016 at 3:37 PM, 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).
Re: [AFMUG] Data center temperatures
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 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 > *Sent:* Wednesday, May 11, 2016 3:51 PM > *To:* af@afmug.com > *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 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 >> lesm...@ecsis.net >> > > > > -- > 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. >
Re: [AFMUG] Data center temperatures
I have a jacket I leave in the data room. Thankfully our noc is a different building from retail. On Wed, May 11, 2016 at 3:51 PM, That One Guy /sarcasm wrote: > 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 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 >> lesm...@ecsis.net > > > > > -- > 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.
Re: [AFMUG] Data center temperatures
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 Sent: Wednesday, May 11, 2016 3:51 PM To: af@afmug.com 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 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 lesm...@ecsis.net -- 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.
Re: [AFMUG] Data center temperatures
62 to 72 is all I was looking for, just curious =) Thanks for the quick answers. Josh Luthman Office: 937-552-2340 Direct: 937-552-2343 1100 Wayne St Suite 1337 Troy, OH 45373 On Wed, May 11, 2016 at 4:52 PM, Eric Kuhnke wrote: > Depends on where the temperature sensors are? If we're talking about > colocation type cabinets with front-to-rear airflow, and mesh doors on both > ends, air intake temperature should be around 20C on the intake side. > > Or open two-post/relay racks in a room? > > Hot aisle/cold aisle separated, or everything together in one space? > > The key metric is air intake temperature, most routers have a separate > SNMP OID for intake temperature where there is a small diode near the air > intake vents. Same with better servers, there's an air intake sensor on > Dells and others near the front intake fans. > http://en.community.dell.com/cfs-file/__key/communityserver-discussions-components-files/956/Fan2.PNG > > > > > On Wed, May 11, 2016 at 1:37 PM, 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). >> > >
Re: [AFMUG] Data center temperatures
Depends on where the temperature sensors are? If we're talking about colocation type cabinets with front-to-rear airflow, and mesh doors on both ends, air intake temperature should be around 20C on the intake side. Or open two-post/relay racks in a room? Hot aisle/cold aisle separated, or everything together in one space? The key metric is air intake temperature, most routers have a separate SNMP OID for intake temperature where there is a small diode near the air intake vents. Same with better servers, there's an air intake sensor on Dells and others near the front intake fans. http://en.community.dell.com/cfs-file/__key/communityserver-discussions-components-files/956/Fan2.PNG On Wed, May 11, 2016 at 1:37 PM, 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). >
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 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 > lesm...@ecsis.net > -- 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.
Re: [AFMUG] Data center temperatures
Ours is at 68deg F, and we monitor dewpoint and humidity ranges. However... http://www.geek.com/chips/googles-most-efficient-data-center-runs-at-95-degrees-1478473/ On Wed, May 11, 2016 at 3:37 PM, 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).
Re: [AFMUG] Data center temperatures
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 lesm...@ecsis.net