Andrew Dunstan <and...@dunslane.net> writes: > There seem to be a bunch of recent failures, and not just on crake, and > not just on HEAD: > <https://buildfarm.postgresql.org/cgi-bin/show_failures.pl?max_days=14&member=&stage=recoveryCheck&filter=Submit>
There were a batch of recovery-stage failures ending about 9 days ago caused by instability of the new 043_vacuum_horizon_floor.pl test. Once you take those out it doesn't look quite so bad. regards, tom lane