On Wed, Dec 2, 2020 at 12:07 PM Tom Lane <t...@sss.pgh.pa.us> wrote: > I'm also wondering a bit why the issue isn't affecting the newer > branches. It's certainly not because we made the test shorter ...
I looked at htop while it was building the 9.x branches and saw pg_basebackup sitting in D state waiting for glacial I/O. Perhaps this was improved by the --no-sync option that got sprinkled around the place in later releases?
