On 2020-Nov-26, Fujii Masao wrote:
> On the second thought, walsender doesn't wait forever unless
> wal_sender_timeout is disabled, even in the case in discussion?
> Or if there is the case where wal_sender_timeout doesn't work expectedly,
> we might need to fix that at first.
Hmm, no, it doesn't wait forever in that sense; tracing with the
debugger shows that the process is looping continuously.
