fping plugin causing growing fping defunct/zombie processes

[mattlan]

vaping version: 1.4.2
pypy version: 7.3.5 (python 3.7.10)
fping version: 3.10

vaping config:

probes:
  - name: latency
    type: std_fping
    output:
      - prometheus
    testing_hosts:
      hosts:
        - host: 1.1.1.1
          name: Cloudflare DNS
plugins:
  - name: std_fping
    type: fping
    count: 10
    interval: 3s
    output:
      - prometheus
  - name: prometheus
    type: prometheus
    port: 9090

As summary notes, when leveraging the fping plugin I am seeing these continually grow over time until the vaping process eventually dies.

root     15904  0.0  0.0      0     0 pts/0    Z    12:39   0:00 [fping] <defunct>
root     15968  0.0  0.0      0     0 pts/0    Z    12:40   0:00 [fping] <defunct>
root     15974  0.0  0.0      0     0 pts/0    Z    12:40   0:00 [fping] <defunct>
root     15979  0.0  0.0      0     0 pts/0    Z    12:40   0:00 [fping] <defunct>
root     15991  0.0  0.0      0     0 pts/0    Z    12:40   0:00 [fping] <defunct>

I believe I've resolved this locally by changing these lines to this:

with self.popen(args, stdout=subprocess.PIPE, stderr=subprocess.STDOUT) as proc:
    for line in iter(proc.stdout.readline, b""):
        data.append(self.parse_verbose(line.decode("utf-8")))

If I get a chance this week will submit a PR, but just noting this in case anyone else runs into this in the meantime.

[vegu]

Thanks for the report - will look at this asap.

[vegu]

@mattlan Not having any luck reproducing on centos 7 with python 3.7.4 or python 3.9.0, also tried with and without --debug flag

Proposed fix looks reasonable though, will discuss with @grizz

[mattlan]

Odd, maybe something specific to pypy? It was happening every time I ran it for me, then went away when I made the referenced change.

[mattlan]

I also have a few internal hosts as part of my testing_hosts, in case that also matters. I omitted them from my attached config.