On Sun, May 28, 2017 at 9:41 AM, Ludovic Vaugeois-Pepin <ludovicvp@xxxxxxxxx> wrote: > On Sun, May 28, 2017 at 9:49 AM, Ludovic Vaugeois-Pepin > <ludovicvp@xxxxxxxxx> wrote: >> It really is set at 3s on all servers (master and standbies) earlier in the "deployment" process at the same time "listen_addresses", "hot_standby", and others are set. This doesn't seem to happen every time I run tests. I increased logging to DEBUG1. This is what I got when the problem occurred again. Note that the empty line is there in the log file. Aren't those logs coming from a standby that is being reconnected to a promoted standby? In short you restarted this standby with a fresh recovery.conf. The WAL receiver would try to fetch history file data for validity checks when it notices that there is a timeline jump, when it needs to itself jump to a new timeline, and the logs you are showing highlight that. -- Michael -- Sent via pgsql-general mailing list (pgsql-general@xxxxxxxxxxxxxx) To make changes to your subscription: http://www.postgresql.org/mailpref/pgsql-general
