On Fri, Mar 30, 2012 at 4:46 PM, Thomas Sattler <[email protected]> wrote:
>>> P.S.: Why do we use 'awk' within 'perl'? >> >> Because it runs on the remote machine, and `awk` has less >> overhead than `perl`. (Though since it's waiting for at >> least 1 second to gather stats, it may not be a >> significant overhead.) > > Perl is not my prefered language so I might be wrong with the > following, but I thought that sub "swap_activity" used 'qx' > to execute "$swap_activity" on the local machine. > > Reading the code again I see something I ignored last time: > 'qx' spawns a background process to get (and save!) these > information. At first I thought it would read them back > into 'parallel'. The reason for this is that Parallel should not wait for the load to be computed, if it for some reason takes a lot of time. So instead it will spawn off a process to do this and read the output back next time. /Ole
