Read mpstat.log file and convert it by aggregating values across NUMA nodes. Read aggregation of CPUs from output of lscpu command.
Usage
./mpstat2node.py --lscpu <(lscpu) < mpstat.log
or
mpstat -P ALL 5 | ./mpstat2node.py --lscpu <(lscpu)
Process mpstat output on the fly for the mpstat process which is already running:
-
mpstat -P ALL 5 > mpstat.txt
-
tail -f -n +1 mpstat.txt | mpstat2node.py --lscpu <(lscpu)
Retrieve mpstat output from another machine:
- Collect mpstat data
lscpu > lscpu.txt
mpstat -P ALL 5 > mpstat.txt
- Aggreagate mpstat data
./mpstat2node.py --lscpu lscpu.txt < mpstat.txt