On Mon, 30 Mar 2020 at 17:57, Laurenz Albe <laurenz.a...@cybertec.at> wrote: > How can it be that even after an explicit VACUUM, this patch can cause > unstable regression test results?
I only added vacuums for mcv_lists. The problem with petalura [1] is with the functional_dependencies table. I'll see if I can come up with some way to do this in a more deterministic way to determine which tables to add vacuums for, rather than waiting for and reacting post-failure. [1] https://buildfarm.postgresql.org/cgi-bin/show_log.pl?nm=petalura&dt=2020-03-30%2002%3A20%3A03
