All block devices are probed synchronously and sequentially on startup before a well known name ownership is requested on the system bus. This can take seconds (or even minutes) on large systems.
The time.sleep(3) in the tests cause system udisks daemon to be autospawned if installed and tests are executed on that one.
Possible solutions:
on systemd-based systems, stop and mask the udisks2.service
optionally kill the binary if found running
watch the well known name presence on the bus, instead of an arbitrary sleep
Issues to solve:
reenable the system service on cleanup, even in case of failure or manual break (Ctrl+C)
(just noting here so I don't forget)
https://github.com/storaged-project/udisks/blob/94e4fef6879f7407fcf7a284cb1c7d7bc71ac3b3/src/tests/dbus-tests/run_tests.py#L217
All block devices are probed synchronously and sequentially on startup before a well known name ownership is requested on the system bus. This can take seconds (or even minutes) on large systems.
The
time.sleep(3)
in the tests cause system udisks daemon to be autospawned if installed and tests are executed on that one.Possible solutions:
udisks2.service
Issues to solve: