Here is a small collection of various improvements for run_all.py script.
The first patch limits the number of tests that can be ran in parallel. This was very useful for us when validating MPTCP packetdrill tests (with a fork version) on a public CI with only 2 cores and a debug kernel config.
The second patch is linked to the first one, simply improving --serialized while at it.
The third patch is useful when someone needs to understand what happened: running packetdrill with -v (or more)
The fourth patch helps capturing packet traces per test for further investigation.
The fifth patch adds a "dry run" mode, simply to print the different commands that are supposed to be executed, e.g. to relaunch specific tests with a different commands.
Here is a small collection of various improvements for
run_all.py
script.The first patch limits the number of tests that can be ran in parallel. This was very useful for us when validating MPTCP packetdrill tests (with a fork version) on a public CI with only 2 cores and a debug kernel config.
The second patch is linked to the first one, simply improving
--serialized
while at it.The third patch is useful when someone needs to understand what happened: running
packetdrill
with-v
(or more)The fourth patch helps capturing packet traces per test for further investigation.
The fifth patch adds a "dry run" mode, simply to print the different commands that are supposed to be executed, e.g. to relaunch specific tests with a different commands.