Developing a software for tracking the memory states of the machines in the LHCb Filter Farm

The LHCb Event Filter Farm consists of more than 1500 server nodes with a total amount of roughly 65 TB operating memory .The memory is crucial for the success of the LHCb experiment, since the proton-proton collisions are temporarily stored on these memory modules. Unfortunately, the aging nodes of the server farm occasionally suffer losses of their memory modules. The lower the available memory, the lower performance we can get out of it. Inducing the users or administrators to pay attention to this matter is inefficient. One needs to upgrade it to an acceptable way. The aim of this project was to develop a software to monitor a set of test machines. The software stores the data of the memory sticks in advance in a database which will be used for future reference. Then it checks the memory sticks at a future time instant to find any failures. In the case of any such losses the software looks up in the database to find out which memory sticks have lost and displays all information of those sticks in a log file. The software is scalable and tested for a couple of test machines.