On Thu, Sep 13, 2018 at 8:20 AM, Michael Paquier <michael@paquier.xyz> wrote:
On Wed, Sep 12, 2018 at 03:55:00PM -0400, Tom Lane wrote: > Although pg_ctl could sneak it in between forking and execing, it seems > like it'd be cleaner to have the postmaster proper do it in response to > a switch that pg_ctl passes it. That avoids depending on the fork/exec > separation, and makes the functionality available for other postmaster > startup mechanisms, and opens the possibility of delaying the detach > to the end of startup.
I would be fine with something happening only once the postmaster thinks that consistency has been reached and is open for business, though I'd still think that this should be controlled by a switch, where the default does what we do now. Feel free to ignore me if I am outvoted ;) -- Michael
From the respective of users instead of developers, I think for such important
service, tty should be dissociated, i.e. postmaster should be daemonized by
default (and even should have default log filename?) or be setsid()-ed at least.
For concerns from developers, maybe a shell alias, or an internal switch in pg_ctl,
or env variable guard in postmaster code could address.
I'm not sure the several-years-ago LINUX_OOM_ADJ issue (to resolve that,
silient_mode was removed) still exists or not. I don't have context about that, so
I conceded to use setsid() even I prefer the deleted silent_mode code.
