I'm the iperf Debian maintainer. I've received this bug report from one user:
I try to measure the network transfer rate between several machines
with a cronjob every 10 minutes. The results are that the measured
network transfer rate decreases constantly. First I was wondering but
when I watched "top" I saw the explanation for this behaviour.
A fresh started server process from iperf needs only few ram and cpu:
PID USER PR NI VIRT RES SHR S %CPU %MEM TIME+ COMMAND
2379 root 20 0 23560 432 292 S 0.0 0.0 0:00.00 iperf
After a few hours the server process of iperf needs a lot of resources:
PID USER PR NI VIRT RES SHR S %CPU %MEM TIME+ COMMAND
6248 root 15 0 574m 1568 436 S 43.6 0.1 226:51.79 iperf
The situation is worse after more than 1 day:
PID USER PR NI VIRT RES SHR S %CPU %MEM TIME+ COMMAND
1241 root 20 0 2032m 5460 420 S 97.0 4.2 2553:28 iperf
The server is started one time with: iperf -sD
The client is started with: iperf -f KBytes -c ServerIPAdresse
Is this a duplicate of [#12] ("Memory Leak in Iperf 2.02")?
Related
Bugs: #12