Randomly haf_node is not close by SIGKILL
Randomly, the node is not being killed by SIGKILL from the close() function in node.py. I noticed this issue on a node with an implemented SQL serializer. The problem occurs when tests are run concurrently or one after another.
Steps to replicate this bug:
- Run all mirrornet tests concurrently or one after another:
tests/integration/system/haf/mirrornet_tests
. - All tests should pass with warnings related to incorrect node closure.
- Some hiveds should still be running. You can verify this by comparing the process IDs (pids) of the running processes with the pids from the test logs. This issue occurs randomly for some tests.
Unfortunately, I am unable to reproduce this error consistently. It seems to happen very rarely.