On Wed, Jun 4, 2014 at 12:14 PM, Greg Stark <stark@mit.edu> wrote:
> Because it had noticed the incomplete
> split it had skipped every restartpoint and the next time I tried to
> start it it insisted on restarting recovery from the beginning.
Incidentally this also caused operational problems. Suppressing all
restartpoints also meant Postgres kept around all the WAL from the
entire recovery. This caused the pg_xlog volume to fill up.
--
greg