Sun System Handbook - ISO 3.4 June 2011 Internal/Partner Edition | |||
|
|
Solution Type Technical Instruction Sure Solution 1012632.1 : Sun Enterprise [TM] E3500/E4500/E6500 Servers: Environmental Best Practices
PreviouslyPublishedAs 217380 Description Description The Enterprise Classic server machines were designed to function in a broad range of environmental conditions. Mean Time Between Failure, (MTBF) statistics were calculated taking in to account servers running in worst case environmental conditions. Historic data has shown that stability can be increased, and the time between failures can be lengthened considerably, if the servers can be consistently run in optimal temperature and humidity environments. Since data center environments are typically designed to run the machines installed in them, at what would be the same temperature and humidity as optimal temperature and humidity range, making adjustments for out-of-range machines in this type of environment, is usually inexpensive, not time-consuming, and well worth the effort. Temperature / Humidity High Temperature reduces the life of components, causes outages due to reduced time margin should a cooling failure occur, and makes it nearly impossible to control humidity. For all these reasons, machines not running in their optimal range need to be addressed. Most data center experts agree, the optimum range for equipment reliability is normally between 21 to 23 degrees Celsius (70 to 74 degrees Fahrenheit). Data analysis has proven, component life and equipment stability is increased, when, equipment is consistently run in this optimal range. An important side benefit is, humidity, which increases the cooling capacity of air and reduces the risk of static damage to components, is much easier to get right. Optimal temperature makes optimal humidity much easier to control. Cooling failures and the shutdown disasters which are associated with them, can be, in many cases avoided. Machines that are running in optical range, can many times avoid automatic overheating shutdown due to a cooling failure, because there generally will be more time to respond to the problem before temperatures reach unacceptable limits. Low Temperature causes contraction, and that can unseat components which were engineered to run in warmer rooms. Once components are unseated, errors occur and server instability results. Equipment placement and incorrect settings in air-handlers, can cause low temperatures in data centers. Low humidity increases the likelihood of static discharge, which is the number one cause of damage to components. It also reduces the cooling capability of the air traveling over components, which causes the components to operate much hotter than they should, creating early-life failures. Generally, low humidity is linked to high temperatures. If the temperature is too high, it is very difficult to reach standard humidity levels. If humidity drops below 40 %, the possibility of failure due to static discharge and overheating of components, doubles. If humidity drops below 30%, it is critically low and equipment failures due to static discharge and component overheating can be expected. Since humidity increases the efficiency of cooling, keeping proper and consistent levels in the data center has the added benefit of saving money. High humidity , (wetness), causes corrosion to conductive surfaces; failures will result. High humidity can also be dangerous, since water is conductive, and it increases the possibility of electrical shock. If humidity is high enough, and it reaches the right temperature, it has been known to rain in data centers, ruining electronic equipment. High humidity is commonly due to misadjusted air handlers, or under floor water leaks. Measurement Basics In order to properly measure temperature and humidity, sufficient logging sensors should be obtained. These are not expensive, and readily available. They should be a part of every data center's basic equipment. They need to be calibrated periodically, so that the data collected is accurate. Equipment Measurements The temperature and humidity of the air at the intake of Servers, Routers, and Switches, is important. Typically, air-intake temperature measurements are taken about 5cm (2 inches) in front of the equipment. The sensor should be in position for several days to get accurate data points, to judge whether or not the temperature and humidity is stable and correct. The intake to all equipment should be between 21 to 23 degrees Celsius (70 to 74 degrees Fahrenheit). The humidity ratings should be 45 to 50% at the intake. Room Measurements When measuring room temperature, it is best practice to measure a large grid at one time, taking into account the elevation of the probes that are placed. If the sensors are not permanent, expect to leave them in place for several days in order to get an historical picture of temperature and humidity stability. Temperature gradient is the difference between the air temperature at the lower levels and the upper levels of a room. The gradient in a room should never exceed more than a few degrees. There is a risk of overheating devices installed higher in the racks, while lower machines are cool. In order to minimize this risk, temperature measurements should always be made at 2 (0.6 meter's) feet and 6 feet (1.8 meter's) off the floor. The high temperature subtracted from the low temperature equals the gradient. If the gradient is greater then 2 degrees, there is an air-pressure problem. It could be caused by dirty airfilters in the airhandler, equipment positioning on the floor, or lack of air-handler capacity in the room. Sensors should be placed to capture gradient information. Sensors should be placed on a grid at least every 20 feet. If a hot/dry spot is observed, it is usually due to a concentration of equipment in a small area, or the equipment is placed incorrectly within the room. Conclusion If Temperature and Humidity are not within specifications, it is well worth the effort and small expense to get it right. A datacenter with a good environment will cost less to run, reduce the risk of production outages, and save money. Always consult an Heating, Ventilating and Air-Conditioning(HVAC) datacenter expert, if questions arise. They can be an invaluable resource when trying to resolve tough problems, and during a move or installation of new equipment. The table below, documents the published environmental temperature and humidity ranges for Sun Enterprise[TM] E3500/E4500/E6500 servers. Steps to Follow Sun Enterprise [TM] E3500/E4500/E6500 Servers: Environmental Best Practices
|
E3500/4500 /5500/6500 Temp and Humidity
|
Non-Operating Low
|
Operating Low
|
Optimal Range
|
Operating High Range
|
Non-Operating High
|
Temperature
|
Below 5oC Below 41oF
|
5oC to 20oC 41oF to 70 oF
|
21o C to 23oC 70oFto73.5oF
|
23oC to 60oC 74oFto104oF
|
Above 61oC Above104oF
|
Relative Humidity, Non-Condensing
|
Below 20 %
|
20% to 45%
|
45% to 50%
|
51% to 80%
|
Above 80% |
Temperature and humidity change:
Severe temperature or relative humidity swings should be avoided. Conditions should not be allowed to change by more than 5.5 C or 10% RH in any 60-minute period of operation.
SEE:
<Document: 1011650.1> for more specific information as to the types of errors which can be incurred by operating a server outside of the optimal range.
Sun Enterprise 6x00/5x00/4x00/ 3x00 Systems Site Planning Guide
For more information about proper installation and system cooling
http://www.sun.com/products-n-solutions/hardware/docs/pdf/805-2817-11.pdf
This solution has no attachment
Copyright © 2011 Sun Microsystems, Inc. All rights reserved.
Feedback |
|||