Wiki run-to-completion example with fio misbehaves

[olivierpierre]

During the analysis, the guided example we have on the wiki with fio misbehaves and somehow spawns a lot of threads hogging 100% of the CPU (see screenshot below). They do not finish: fio is configured to run for 2s max, and loupe to kill misbehaving instances after 30 seconds.

[olivierpierre]

I will switch the run to completion application example in the wiki from fio to sysbench to avoid that issue. In absolute this does not solve the problem with fio. To reproduce it, use the Dockerfile and test script below.

Dockerfile.fio:

FROM loupe-base:latest

# We are going to build fio from source, install dependencies to build it
RUN apt update
RUN apt install -y build-essential git

# Clone fio's repo:
RUN git clone https://github.com/axboe/fio.git

# Build fio
RUN cd fio && make -j

# fio is ready, we still need to copy the test script in the container
COPY dockerfile_data/fio-test.py /root/fio-test.py
RUN chmod a+x /root/fio-test.py

# And finally the invocation of Loupe within the container it will run in:
CMD /root/explore.py --output-csv -t /root/fio-test.py --timeout 30 \
		     -b /root/fio/fio --test-sequential --disable-static \
		     -- --name=test --size=4K --time_based --runtime=2s

fio-test.py:

#!/usr/bin/python3

import sys, re

regex=re.compile("READ: bw=([0-9]+)") # We'll use this regular expression to grab the bandwidth

output = ""
with open (sys.argv[1], "r") as f: # The first command line parameter points to a file
    output = f.read()              # that contains the output of the last test iteration

# If our regular expression matches, grab the bandwidth and check that it's > 0
match = re.search(regex, output)
if match:
    bandwidth = int(match.groups()[0])
    if bandwidth > 0:
        exit(0)        # Bandwith is positive, test suceeded

exit(1) # Couldn't match the bandwidth regex, or it's not positive, test failed