I'll be looking into why we're getting critical swap alerts, and I'll find something like:
PID USER PR NI VIRT RES SHR S %CPU %MEM TIME+ SWAP COMMAND
14683 user1 30 10 16.0g 772m 744 S 0.3 0.3 277:24.87 6.5g MATLAB
14576 user1 30 10 8125m 1.1g 736 S 0.3 0.4 261:16.73 4.1g MATLAB
where the TIME+ column is frozen for those processes.
My question is, in the above, what does this imply?
As I understand it, stopped time means that it's a dormant process/not being processed by the CPU. However the above example shows 0.3% CPU use, so despite the fact that it's very small, shouldn't the time column continue incrementing?
Are these processes "finished"? In which case, how are they cleaned up? Does the user who started them need to acknowledge them or something for those resources to be freed up?
Best Answer
If you run a fake job such as
sleep 120
and then watch it inhtop
you'll notice that its state isS
aka "SLEEP" and the processesTIME
remains at 0:00.00 for the duration.That's because that process is consuming 0 CPU time, which is the intent of the
TIME
column. It tracks the amount of CPU time a given process has used.This count can be confusing at times if a process can run on multiple CPU cores, since it can seemingly consume more time than is available. All that's going on there is that if you have X number of cores, then your time can show up as X * TIME.