On Thu, Aug 2, 2018 at 10:30 PM, Tom Lane <t...@sss.pgh.pa.us> wrote:
> Paul Guo <p...@pivotal.io> writes: > > [ make the postmaster execute setsid() too ] > > I'm a bit skeptical of this proposal. Forcing the postmaster to > dissociate from its controlling terminal is a good thing in some > scenarios, but probably less good in others, and manual postmaster > starts are probably mostly in the latter class. > > I wonder whether having "pg_ctl start" do a setsid() would accomplish > the same result with less chance of breaking debugging usages. > > regards, tom lane > Yes, if considering the case of starting postmaster manually, we can not create a new session in postmaster, so pg_ctl seems to be a good place for setsid() call. Attached a newer patch. Thanks.
0001-Create-a-new-session-for-postmaster-in-pg_ctl-by-cal.patch
Description: Binary data