- Create
diag_netstat.sh
:
#!/bin/sh
OUTPUTFILE="diag_netstat_$(hostname)_$(date +"%Y%m%d_%H%M%S").log"
INTERVAL="${1:-30}" # First argument or a default of 30 seconds
while true; do
echo "diag: $(date +"%Y%m%d %H%M%S %N %Z") iteration" >> "${OUTPUTFILE}" 2>&1
netstat -antop >> "${OUTPUTFILE}" 2>&1
sleep ${INTERVAL}
done
- Make it executable:
chmod +x diag_netstat.sh
- Start it (default interval of 30 seconds or specify an alternate
interval as the first argument):
nohup ./diag_netstat.sh &
- Reproduce the problem
- If at any point you need to reduce this disk usage of the output
file during the test, truncate the file:
cat /dev/null > diag_netstat*log
- Stop the script:
pkill -f diag_netstat
- Upload
diag_netstat_*
and nohup.out
Previous Section (Linux perf Recipe) |
Next Section (Linux basics Recipe) |
Back to Table of Contents