Open espresso-ci opened 1 year ago
The virtual_sites_relative.py
test case times out at random. We recently increased the verbosity of the test to check which function was responsible for the time out, however it doesn't happen every time in the same function.
Here it happens after the test_pos_vel_forces
teardown and before the test_vs_exceptions
setup:
82/186 Test #94: virtual_sites_relative ........................................***Timeout 300.01 sec
test_aa_method_switching (__main__.VirtualSites) ... ok
test_lj (__main__.VirtualSites)
Run LJ fluid test for different cell systems. ... ok
test_pos_vel_forces (__main__.VirtualSites) ... ok
Here it happens right when test_lj
starts, before the class can print the method docstring:
81/185 Test #94: virtual_sites_relative ........................................***Timeout 300.01 sec
test_aa_method_switching (__main__.VirtualSites) ... ok
test_lj (__main__.VirtualSites)
New data point:
82/185 Test #94: virtual_sites_relative ........................................***Timeout 300.01 sec
test_aa_method_switching (__main__.VirtualSites) ... setUp(Fri Dec 23 20:01:26 2022) tearDown(Fri Dec 23 20:01:26 2022) ok
test_lj (__main__.VirtualSites)
The deadlock happens before entering the setUp
method of the LJ test.
Also, the bug seems more likely to happen when both the default
and maxset
CI jobs are running simultaneously on the coyote11
runner.
https://gitlab.icp.uni-stuttgart.de/espressomd/espresso/pipelines/19197