target low aisle temps in the datacenter
Now that I've started monitoring ipmi temperature sensors, I'm starting to see some systems that are reporting low temperature alarms. Most of the systems I've looked at alarm at 15 degrees C / 59F. The spec sheets for most of the servers, and the tape cartridges have the critical low operating temperature at 50F. I think we're currently very close to that at periods during the day, particularly in front of the tape library. I had a below 12 degree C alarm on tape library frame 2 at 3:11pm on the 8th.
From reading the ASHRE standard documents it looks like we're below spec for the low temp range for the 2004 (68F) and the 2008 (65F) version. Both of them have specific comments that tapes have more stringent class 1 temperature range requirements. It looks like we're at least 10 degrees F below the maximum inlet of 77F (2004) or 80F (2008) in the warmest cold aisles of the room.
It looks like we could use a 5 degree F bump on some of the cold aisle targets, does anyone know if it would be possible to get a higher setpoint in a few strategic areas? Mainly the cold aisle that runs in front of the tape library and in front of NMPDR. ---- Daniel Murphy-Olson Systems Administrator Mathematics & Computer Science Division Argonne National Laboratory 630-252-0055
Not without some engineering (ie baffles). I'll see where we are on getting that. -- Craig (from my mobile) Dan Olson <[email protected]> wrote: Now that I've started monitoring ipmi temperature sensors, I'm starting to see some systems that are reporting low temperature alarms. Most of the systems I've looked at alarm at 15 degrees C / 59F. The spec sheets for most of the servers, and the tape cartridges have the critical low operating temperature at 50F. I think we're currently very close to that at periods during the day, particularly in front of the tape library. I had a below 12 degree C alarm on tape library frame 2 at 3:11pm on the 8th.
From reading the ASHRE standard documents it looks like we're below spec for the low temp range for the 2004 (68F) and the 2008 (65F) version. Both of them have specific comments that tapes have more stringent class 1 temperature range requirements. It looks like we're at least 10 degrees F below the maximum inlet of 77F (2004) or 80F (2008) in the warmest cold aisles of the room.
It looks like we could use a 5 degree F bump on some of the cold aisle targets, does anyone know if it would be possible to get a higher setpoint in a few strategic areas? Mainly the cold aisle that runs in front of the tape library and in front of NMPDR. ---- Daniel Murphy-Olson Systems Administrator Mathematics & Computer Science Division Argonne National Laboratory 630-252-0055
We're probably driving the cold as much as we can to try and deal with the hot spots in the room. What about simply replacing some of the vented tiles near the library with solid ones? -- Hunter Matthews Unix Administrator Office: Bldg 221 Room B240 Argonne National Labs, MCS Key: F0F88438 / FFB5 34C0 B350 99A4 BB02 9779 A5DB 8B09 F0F8 8438 Never take candy from strangers. Especially on the internet. On Aug 11, 2012, at 11:46 AM, Dan Olson wrote:
Now that I've started monitoring ipmi temperature sensors, I'm starting to see some systems that are reporting low temperature alarms. Most of the systems I've looked at alarm at 15 degrees C / 59F. The spec sheets for most of the servers, and the tape cartridges have the critical low operating temperature at 50F. I think we're currently very close to that at periods during the day, particularly in front of the tape library. I had a below 12 degree C alarm on tape library frame 2 at 3:11pm on the 8th.
From reading the ASHRE standard documents it looks like we're below spec for the low temp range for the 2004 (68F) and the 2008 (65F) version. Both of them have specific comments that tapes have more stringent class 1 temperature range requirements. It looks like we're at least 10 degrees F below the maximum inlet of 77F (2004) or 80F (2008) in the warmest cold aisles of the room.
It looks like we could use a 5 degree F bump on some of the cold aisle targets, does anyone know if it would be possible to get a higher setpoint in a few strategic areas? Mainly the cold aisle that runs in front of the tape library and in front of NMPDR.
---- Daniel Murphy-Olson Systems Administrator Mathematics & Computer Science Division Argonne National Laboratory 630-252-0055
We did replace some of the perf tiles with solid ones, however with less flow we create more variance in the library. The operating range on the tapes / drives is larger at a higher CFM. I don't think it is low to deal with hot spots. The hot areas in the room are pretty cool. delta T across most of the servers is 25 deg F, so we don't have any super heat producers. I think 5 or 10 deg F additional supply temp wouldn't hurt anything that is visible on our temp monitors. I think the common way to do this is mix return air with the supply air, there should be a damper somewhere in the system to control that. >50deg F supply air that is coming from the outlets sounds like we're not mixing return air at all. ---- Daniel Murphy-Olson Systems Administrator Mathematics & Computer Science Division Argonne National Laboratory 630-252-0055 ----- Original Message ----- From: "Hunter Matthews" <[email protected]> To: "Dan Olson" <[email protected]> Cc: "Core Admins" <[email protected]> Sent: Saturday, August 11, 2012 3:43:41 PM Subject: Re: target low aisle temps in the datacenter We're probably driving the cold as much as we can to try and deal with the hot spots in the room. What about simply replacing some of the vented tiles near the library with solid ones? -- Hunter Matthews Unix Administrator Office: Bldg 221 Room B240 Argonne National Labs, MCS Key: F0F88438 / FFB5 34C0 B350 99A4 BB02 9779 A5DB 8B09 F0F8 8438 Never take candy from strangers. Especially on the internet. On Aug 11, 2012, at 11:46 AM, Dan Olson wrote:
Now that I've started monitoring ipmi temperature sensors, I'm starting to see some systems that are reporting low temperature alarms. Most of the systems I've looked at alarm at 15 degrees C / 59F. The spec sheets for most of the servers, and the tape cartridges have the critical low operating temperature at 50F. I think we're currently very close to that at periods during the day, particularly in front of the tape library. I had a below 12 degree C alarm on tape library frame 2 at 3:11pm on the 8th.
From reading the ASHRE standard documents it looks like we're below spec for the low temp range for the 2004 (68F) and the 2008 (65F) version. Both of them have specific comments that tapes have more stringent class 1 temperature range requirements. It looks like we're at least 10 degrees F below the maximum inlet of 77F (2004) or 80F (2008) in the warmest cold aisles of the room.
It looks like we could use a 5 degree F bump on some of the cold aisle targets, does anyone know if it would be possible to get a higher setpoint in a few strategic areas? Mainly the cold aisle that runs in front of the tape library and in front of NMPDR.
---- Daniel Murphy-Olson Systems Administrator Mathematics & Computer Science Division Argonne National Laboratory 630-252-0055
That > should have be a < 50deg F ---- Daniel Murphy-Olson Systems Administrator Mathematics & Computer Science Division Argonne National Laboratory 630-252-0055 ----- Original Message ----- From: "Dan Olson" <[email protected]> To: "Hunter Matthews" <[email protected]> Cc: "Core Admins" <[email protected]> Sent: Saturday, August 11, 2012 4:19:43 PM Subject: Re: target low aisle temps in the datacenter We did replace some of the perf tiles with solid ones, however with less flow we create more variance in the library. The operating range on the tapes / drives is larger at a higher CFM. I don't think it is low to deal with hot spots. The hot areas in the room are pretty cool. delta T across most of the servers is 25 deg F, so we don't have any super heat producers. I think 5 or 10 deg F additional supply temp wouldn't hurt anything that is visible on our temp monitors. I think the common way to do this is mix return air with the supply air, there should be a damper somewhere in the system to control that. >50deg F supply air that is coming from the outlets sounds like we're not mixing return air at all. ---- Daniel Murphy-Olson Systems Administrator Mathematics & Computer Science Division Argonne National Laboratory 630-252-0055 ----- Original Message ----- From: "Hunter Matthews" <[email protected]> To: "Dan Olson" <[email protected]> Cc: "Core Admins" <[email protected]> Sent: Saturday, August 11, 2012 3:43:41 PM Subject: Re: target low aisle temps in the datacenter We're probably driving the cold as much as we can to try and deal with the hot spots in the room. What about simply replacing some of the vented tiles near the library with solid ones? -- Hunter Matthews Unix Administrator Office: Bldg 221 Room B240 Argonne National Labs, MCS Key: F0F88438 / FFB5 34C0 B350 99A4 BB02 9779 A5DB 8B09 F0F8 8438 Never take candy from strangers. Especially on the internet. On Aug 11, 2012, at 11:46 AM, Dan Olson wrote:
Now that I've started monitoring ipmi temperature sensors, I'm starting to see some systems that are reporting low temperature alarms. Most of the systems I've looked at alarm at 15 degrees C / 59F. The spec sheets for most of the servers, and the tape cartridges have the critical low operating temperature at 50F. I think we're currently very close to that at periods during the day, particularly in front of the tape library. I had a below 12 degree C alarm on tape library frame 2 at 3:11pm on the 8th.
From reading the ASHRE standard documents it looks like we're below spec for the low temp range for the 2004 (68F) and the 2008 (65F) version. Both of them have specific comments that tapes have more stringent class 1 temperature range requirements. It looks like we're at least 10 degrees F below the maximum inlet of 77F (2004) or 80F (2008) in the warmest cold aisles of the room.
It looks like we could use a 5 degree F bump on some of the cold aisle targets, does anyone know if it would be possible to get a higher setpoint in a few strategic areas? Mainly the cold aisle that runs in front of the tape library and in front of NMPDR.
---- Daniel Murphy-Olson Systems Administrator Mathematics & Computer Science Division Argonne National Laboratory 630-252-0055
I< understood <=> you.
-- Hunter Matthews Unix Administrator Office: Bldg 221 Room B240 Argonne National Labs, MCS Key: F0F88438 / FFB5 34C0 B350 99A4 BB02 9779 A5DB 8B09 F0F8 8438 Never take candy from strangers. Especially on the internet. On Aug 11, 2012, at 4:21 PM, Dan Olson wrote:
That > should have be a < 50deg F
---- Daniel Murphy-Olson Systems Administrator Mathematics & Computer Science Division Argonne National Laboratory 630-252-0055
----- Original Message ----- From: "Dan Olson" <[email protected]> To: "Hunter Matthews" <[email protected]> Cc: "Core Admins" <[email protected]> Sent: Saturday, August 11, 2012 4:19:43 PM Subject: Re: target low aisle temps in the datacenter
We did replace some of the perf tiles with solid ones, however with less flow we create more variance in the library. The operating range on the tapes / drives is larger at a higher CFM.
I don't think it is low to deal with hot spots. The hot areas in the room are pretty cool. delta T across most of the servers is 25 deg F, so we don't have any super heat producers. I think 5 or 10 deg F additional supply temp wouldn't hurt anything that is visible on our temp monitors. I think the common way to do this is mix return air with the supply air, there should be a damper somewhere in the system to control that. >50deg F supply air that is coming from the outlets sounds like we're not mixing return air at all.
---- Daniel Murphy-Olson Systems Administrator Mathematics & Computer Science Division Argonne National Laboratory 630-252-0055
----- Original Message ----- From: "Hunter Matthews" <[email protected]> To: "Dan Olson" <[email protected]> Cc: "Core Admins" <[email protected]> Sent: Saturday, August 11, 2012 3:43:41 PM Subject: Re: target low aisle temps in the datacenter
We're probably driving the cold as much as we can to try and deal with the hot spots in the room.
What about simply replacing some of the vented tiles near the library with solid ones?
-- Hunter Matthews Unix Administrator Office: Bldg 221 Room B240 Argonne National Labs, MCS Key: F0F88438 / FFB5 34C0 B350 99A4 BB02 9779 A5DB 8B09 F0F8 8438 Never take candy from strangers. Especially on the internet.
On Aug 11, 2012, at 11:46 AM, Dan Olson wrote:
Now that I've started monitoring ipmi temperature sensors, I'm starting to see some systems that are reporting low temperature alarms. Most of the systems I've looked at alarm at 15 degrees C / 59F. The spec sheets for most of the servers, and the tape cartridges have the critical low operating temperature at 50F. I think we're currently very close to that at periods during the day, particularly in front of the tape library. I had a below 12 degree C alarm on tape library frame 2 at 3:11pm on the 8th.
From reading the ASHRE standard documents it looks like we're below spec for the low temp range for the 2004 (68F) and the 2008 (65F) version. Both of them have specific comments that tapes have more stringent class 1 temperature range requirements. It looks like we're at least 10 degrees F below the maximum inlet of 77F (2004) or 80F (2008) in the warmest cold aisles of the room.
It looks like we could use a 5 degree F bump on some of the cold aisle targets, does anyone know if it would be possible to get a higher setpoint in a few strategic areas? Mainly the cold aisle that runs in front of the tape library and in front of NMPDR.
---- Daniel Murphy-Olson Systems Administrator Mathematics & Computer Science Division Argonne National Laboratory 630-252-0055
When we were mixing, it was getting too hot near fusion and pads. -- Craig On Aug 11, 2012, at 4:19 PM, Dan Olson <[email protected]> wrote:
We did replace some of the perf tiles with solid ones, however with less flow we create more variance in the library. The operating range on the tapes / drives is larger at a higher CFM.
I don't think it is low to deal with hot spots. The hot areas in the room are pretty cool. delta T across most of the servers is 25 deg F, so we don't have any super heat producers. I think 5 or 10 deg F additional supply temp wouldn't hurt anything that is visible on our temp monitors. I think the common way to do this is mix return air with the supply air, there should be a damper somewhere in the system to control that. >50deg F supply air that is coming from the outlets sounds like we're not mixing return air at all.
---- Daniel Murphy-Olson Systems Administrator Mathematics & Computer Science Division Argonne National Laboratory 630-252-0055
----- Original Message ----- From: "Hunter Matthews" <[email protected]> To: "Dan Olson" <[email protected]> Cc: "Core Admins" <[email protected]> Sent: Saturday, August 11, 2012 3:43:41 PM Subject: Re: target low aisle temps in the datacenter
We're probably driving the cold as much as we can to try and deal with the hot spots in the room.
What about simply replacing some of the vented tiles near the library with solid ones?
-- Hunter Matthews Unix Administrator Office: Bldg 221 Room B240 Argonne National Labs, MCS Key: F0F88438 / FFB5 34C0 B350 99A4 BB02 9779 A5DB 8B09 F0F8 8438 Never take candy from strangers. Especially on the internet.
On Aug 11, 2012, at 11:46 AM, Dan Olson wrote:
Now that I've started monitoring ipmi temperature sensors, I'm starting to see some systems that are reporting low temperature alarms. Most of the systems I've looked at alarm at 15 degrees C / 59F. The spec sheets for most of the servers, and the tape cartridges have the critical low operating temperature at 50F. I think we're currently very close to that at periods during the day, particularly in front of the tape library. I had a below 12 degree C alarm on tape library frame 2 at 3:11pm on the 8th.
From reading the ASHRE standard documents it looks like we're below spec for the low temp range for the 2004 (68F) and the 2008 (65F) version. Both of them have specific comments that tapes have more stringent class 1 temperature range requirements. It looks like we're at least 10 degrees F below the maximum inlet of 77F (2004) or 80F (2008) in the warmest cold aisles of the room.
It looks like we could use a 5 degree F bump on some of the cold aisle targets, does anyone know if it would be possible to get a higher setpoint in a few strategic areas? Mainly the cold aisle that runs in front of the tape library and in front of NMPDR.
---- Daniel Murphy-Olson Systems Administrator Mathematics & Computer Science Division Argonne National Laboratory 630-252-0055
participants (3)
-
Craig Stacey -
Dan Olson -
Hunter Matthews