One of the critical aspects of managing a Linux Server is the regular inspection of the server for heat management issues. Over time, all servers collect dust, etc. that reduces the heat management capabilities of a server. This is especially true with Linux Servers as they will actually run for years without a need to reboot. This article contains some tips on what checks to make when your Linux Server is down for maintenance.

The most important location to check for heat management is the CPU. In this example there are two Xeon processors that are seriously in need to a cleaning. Before you start any cleaning issues be sure to disconnect all power to the box. No ..not just turn it off, but disconnect it as modern motherboards will maintain power to the motherboard even when the server is turned off. The next thing to watch for is to be sure to ground yourself and not create static electricity by moving around and then touching the server. Static electricity is your worst enemy.

Back to the Xeons...one is about 50% impaired in terms of air flow and the other is about 85% impaired. This is a serious issue, especially with Xeon processors that already run hot. You will often be able to tell that this is a problem by monitoring your motherboard BIOS which may tell you CPU temperature or simply be the increased whine of the fans trying to keep up.

Clean the vents with canned air, not an air compressor as the air compressor may add moisture to the air. The canned air should always be held upright so that fluid does not spray on the processor. As you look down into the processors you can see a layer of dust all through the system.

Here is the same Xeons which have been cleaned. Note that they look shiny new, you need to get all that you can out of those vents so that you have optimum air flow.

Also clean all of the fins that are on the motherboard for cooling. These fins on the motherboard will help to maintain stability as the temperature is within the limits of the electronic parts. Note how this looks shiny new...get it clean!

Here is an illustration of the dust that is covering cables on this same box before it was cleaned. These SCSI cables are covered in dust and you can see the plastic cover over this cable has been destroyed because of heat. The plastic is just falling apart, another problem.

You will find that it is also critical to clean the RAM on the motherboard. This is an area that is often overlooked when you think about heat but your RAM chips must also be cool.

You should probably shut down your Linux server every three months to go through the process of cleaning just so you can maintain heat management.

Take a class to learn how to manage Linux Servers!

Live Class Descriptions:
Linux Server Management - course focus on running a Linux Server correctly (Ubuntu 8.04 or CentOS 5), see class outline below.
Linux Server Daemons - course focus on step-by-step install of servers (Ubuntu 8.04 or CentOS 5), installation and management of web servers, DNS servers, mail servers, DHCP,SSH, FTP,etc.

Self-Teach Packages:
Linux Server Management - course focus on running a Linux Server correctly (Ubuntu 8.04 or CentOS 5), see class outline below.
Linux Server Daemons - course focus on step-by-step install of servers (Ubuntu 8.04 or CentOS 5), installation and management of web servers, DNS servers, mail servers, DHCP,SSH, FTP,etc.