On Fri, Jan 28, 2022 at 9:27 AM Andrew Dunstan <and...@dunslane.net> wrote: > I have noticed on a different setup that this test adds 11 minutes to > the runtime of the recovery tests, effectively doubling it. The doubling > is roughly true on faster setups, too. At least I would like a simple > way to disable the test.
Ouch, that's ... a lot. Some randomly selected times: ~20 seconds (dev machines), ~40 seconds (Cirrus CI's Windows image), ~2-3 minutes (very cheap cloud host accounts), ~3 minutes (my Rapsberry Pi pinned onto two CPU cores), ~11 minutes (your Windows number). It would be good to understand why that's such an outlier. Re skipping, I've also been wondering about an exclusion list to skip parts of the regression tests that don't really add recovery coverage but take non-trivial time, like the join tests.
