Do CSG routinely kill processes which run for days and days?
Only if they come to our attention (eg: we receive a complaint about a process or we spot a process is taking a very large amount of CPU time). We try to make a judgement about a process as to whether it has gone “rogue” before killing it. If we are not sure about a particular process and it looks like the process might be connected with coursework or project work than we might reduce its priority and/or suspend it rather than killing it outright. We are more likely to stop processes which are impeding the performance of the shell servers (shell1, shell2, shell3 and shell4) since these are intended for routine remote login by many users in the department.
