On Thu, Dec 21, 2017 at 02:46:15PM +0100, Magnus Hagander wrote: > On Thu, Dec 21, 2017 at 1:38 AM, Michael Paquier <michael.paquier@gmail.com> > wrote: >> I think that the call to pgstat_report_activity in WalSndLoop() should >> be kept as well. There is a small gap between the moment the process >> is started and the first replication command is run. >> > > Eh. But WalSndLoop() is called *after* exec_replication_command(), isn't > it? exec_replication_command() is called from PostgresMain(), and then > calls WalSndLoop(). > > So I agree there is a small gap, but actually moving it to > exec_replication_command() makes that gap smaller than it was before, no?
My turn to read things wrong then, thinking that WalSndLoop() was the main routine used for starting the WAL sender process. You are right removing the call there is adapted.
Could you update the patch?
I thought I had, but I can see now that email was a figment of my imagination :)