Common Server Rack Problems
There are three common problems when it comes to server/mount racks for servers and hardware and they are deeper servers, additional flanges and round holes.
Computer servers are getting deeper and deeper because manufactures are unable to increase their size in any other way. The server would not fit into a server rack if the manufacturer were to make the item taller or wider, so they have to make them deeper. This means that every time they pack it full of technology, that they have to make it a little deeper. This is okay in principle, but it means that servers and hardware are still outgrowing their racks. They are not getting taller or wider but they are making it so that the back door of the cabinets is unable to close. They are selling racks/mounts as deep as 44 inches now, in order to cope with the way servers and hardware are getting deeper. The standard EIA specification for racks does not really address what a manufacturer should do between the rear and front mounting posts. This means that there are sometimes additional mounting features or flanges which often end up colliding with an OEM slide rail and hence prevent installation. You should probably buy a rack that has square holes because the round ones are often incompatible with the server you install. You can add a thread with a cage nut if needs be. If you buy a rack with round holes then you are going to have a problem installing inside of it. On the bright side however, there is research that suggests that severs are not as affected by heat as we initially imagined. Modern servers may be the reason behind this. The University of Toronto did a study around how heat affects servers, by using data from thousands of equipment failures.
Their studies showed that there were often other factors involved with system failures and that sometimes the overheated severs simply added to these problems, instead of causing them. They found that when the heating of servers was a factor, that the failure and problem rate grew in a linear pattern instead of an exponential one. In short, this means that even if your servers are over heated, this does not mean that they are definitely going to fail. It also means that although they are going to slow down, they may not actually fail. If this research is built upon and found to be accurate, then data centres may be able to save a little money on their air conditioning bills, or maybe even run a few more machines in their already large facilities. The assumption that higher heat may cause system failures is a commonly held belief, mainly due to the fact that slightly elevated temperature in server will decrease server efficiency.
Computer servers are getting deeper and deeper because manufactures are unable to increase their size in any other way. The server would not fit into a server rack if the manufacturer were to make the item taller or wider, so they have to make them deeper. This means that every time they pack it full of technology, that they have to make it a little deeper. This is okay in principle, but it means that servers and hardware are still outgrowing their racks. They are not getting taller or wider but they are making it so that the back door of the cabinets is unable to close. They are selling racks/mounts as deep as 44 inches now, in order to cope with the way servers and hardware are getting deeper. The standard EIA specification for racks does not really address what a manufacturer should do between the rear and front mounting posts. This means that there are sometimes additional mounting features or flanges which often end up colliding with an OEM slide rail and hence prevent installation. You should probably buy a rack that has square holes because the round ones are often incompatible with the server you install. You can add a thread with a cage nut if needs be. If you buy a rack with round holes then you are going to have a problem installing inside of it. On the bright side however, there is research that suggests that severs are not as affected by heat as we initially imagined. Modern servers may be the reason behind this. The University of Toronto did a study around how heat affects servers, by using data from thousands of equipment failures.
Their studies showed that there were often other factors involved with system failures and that sometimes the overheated severs simply added to these problems, instead of causing them. They found that when the heating of servers was a factor, that the failure and problem rate grew in a linear pattern instead of an exponential one. In short, this means that even if your servers are over heated, this does not mean that they are definitely going to fail. It also means that although they are going to slow down, they may not actually fail. If this research is built upon and found to be accurate, then data centres may be able to save a little money on their air conditioning bills, or maybe even run a few more machines in their already large facilities. The assumption that higher heat may cause system failures is a commonly held belief, mainly due to the fact that slightly elevated temperature in server will decrease server efficiency.
Source...