rack of A-class servers getting hot.

[someone_4]

Hello
we have a rack of 10 A-class (A-500 9000) servers. We have had them up for a while but we just found out (though emails that I am not looking at)that they are getting event #7 event witch means that they are not getting cooled down. Hp came and told us there was nothing they could do that our equipment room is not cold enough. But have allot of equipment in there everything from a Nortel DMS 250 to a XP-256 and other N and L classes. And they dont complain. Has anyone had this problem? Or does anyone have any advice?

Richard

[Christopher Caldwell · Accepted solution]

We got the same message, under similar circumstances.

According to the ITRC, the message is erroneous and can be fixed with a firware upgrade for the A500.

>------------ Event Monitoring Service Event Notification ------------<

Notification Time: Sat Apr 28 00:46:43 2001

ike sent Event Monitor notification information:

/system/events/core_hw/core_hw is >= 3.
Its current value is MAJORWARNING(3).



Event data from monitor:

Event Time..........: Sat Apr 28 00:46:43 2001
Severity............: MAJORWARNING
Monitor.............: dm_core_hw
Event #.............: 7
System..............: A.B.com

Summary:
Insufficient cabinet fans


Description of Error:

An insufficient number of fans are installed in the cabinet to cool the
unit.

Probable Cause / Recommended Action:

One or more fans have been removed from the cabinet. Replace the fan(s).

Additional Event Data:
System IP Address...: 1.2.3.4
Event Id............: 0x3aea4b3300000000
Monitor Version.....: B.01.00
Event Class.........: System
Client Configuration File...........:
/var/stm/config/tools/monitor/default_dm_core_hw.clcfg
Client Configuration File Version...: A.01.00
Qualification criteria met.
Number of events..: 1
Associated OS error log entry id(s):
None
Additional System Data:
System Model Number.............: 9000/800
EMS Version.....................: A.03.20
STM Version.....................: A.22.00
Latest information on this event:
http://docs.hp.com/hpux/content/hardware/ems/dm_core_hw.htm#7

v-v-v-v-v-v-v-v-v-v-v-v-v D E T A I L S v-v-v-v-v-v-v-v-v-v-v-v-v


Device identification information:

Number of fans present.: 4
Location of fan(s).....: Cabinet
Present fan number(s)..: 0,1,2,3


>---------- End Event Monitoring Service Event Notification ----------<

[someone_4]

Are you boxes warm?
What we did was pop of the front cover and you can feel the air in the back going though them allot better.

Richard

[Keith Hickox]

How tall is the cabinet? Are all 10 boxes racked next to each other, if so, can you spread them out in the rack (ie leave a 2U space).

[Christopher Caldwell]

Our boxes really didn't feel any warmer than any other boxes, that's why we couldn't make sense of the error message (everything seemed to like our temperature).

[David Lieberman_1]

Richard,

Christopher probably has the correct answer for you.

10 A500 (550MHz) servers put out about 2048 BTUs each - about 1/5 of an N4000. At 20,480 BTUs, you're probably not overloading the room AC. HP sells a rack with 41 of them enclosed.

However take a moment and check for compliancy with the operating requirements for the servers:

Pg. 2-6 of the Nov. 2000 HP 9000 Enterprise Servers COnfiguration Guide has the operating temperature and humidity requirements for your A500 servers:

+5 to +35 deg. C (+41 to +95 F) @ <=5000 ft.
15% to 80% RH

This measurement should take place in the vicinity of the rack. Although the computer room may be cool overall, I've seen people aim racks of server's exhaust towards the intake of another's with not-so-suprising results. (Doh!)

If you have the book, you'll see that different servers have different operating environment temperature requirements. So, it is possible (but not very likely) that one piece of gear could overheat next to one that is chugging away just fine.

After taking the temperature and humidity, it's probably worth another call to your CE to have the firmware upgraded etc. At least you'll have the numbers to prove to them that they need to fix something.

Best of luck,

David Lieberman

[someone_4]

Chris ..
Can you tell me the link where it mentions about the error on the A Class being erronous?

Richard

[Dave Wherry]

How is the air conditioning piped through your room? Our datacenter started as a little PC equipment room. The air blows out of the top of the unit and that was fine for those little boxes.

When we started adding larger systems we developed several hot spots in the room. The larger systems (HP, others) draw air from the bottom of the cabinet/rack and exhaust it out the top, except for the V class which is front to back. The best way to keep these systems cool is to pump the cold air under the floor and let the cabinets/racks draw it up through ventilated floor tiles. We exhaust the warm air up into the cooled air, thus warming it and making the AC unit work that much harder. So we have a wind chill near the AC unit and the far end of the room is nearly tropical.

Also, cabinet doors should remain closed. The cabinets are designed to draw the air through and keep the equipment cooled. One place I worked at we had a full time engineer who did thermal studies on each configuration we sold. Leaving the door open defeats the air flow design. That being said, if there is not enough cool air at the bottom of the cabinet to be drawn up, you may need to keep the doors open. The equipment design goes hand in hand with the room design.

We had a product demo at the local HP office a few months ago and toured their demo datacenter. Our guy who built our datacenter was, excuse the pun, blown away by how uniformly cool the room was. He knelt down and felt the air coming through one of the ventilated tiles. I've been asking for that for two years. Instead we have about 5 floor fans around the room moving the hot air.

[Christopher Caldwell]

I logged a call with the ITRC (I didn't find the info on the ITRC website).

[someone_4]

we just did some testing. Yesterday we took the temp. with an electric eye thermometer. And the back of each unit was about 95 degress.
We took the face off one of the units and we checked it today it it went from aound 95 to 80 to 83. This really is an issue whent he face causes 10 to 15 degress difference.

[Bill McNAMARA_1]

I'm getting the same pb on 2 A500s.
The L2k right beside are hot too.
The fans are working fine, but the heat they're trying to dissapate is as hot as my hair dryer.
There is a 4 port 10/100 card in both ncloser to the heat source which seems to be hotter on the turbo slot side than on the cpu side.

I get no errors from gsp or bch but my a's are randomly disconnecting my x sessions. Nothing in logs about the reboot, an ser pim
gives the following see attachment..