On Mon, Jul 23, 2018 at 12:56 PM, Michael Paquier <mich...@paquier.xyz> wrote: > On Mon, Jul 23, 2018 at 01:31:57AM +0000, Tsunakawa, Takayuki wrote: >> First, as Hari-san said, starting the server with "pg_ctl start" on >> the command prompt does not feel like production use but like test >> environment, and that usage seems mostly interactive. Second, the >> current PostgreSQL is not exempt from the same problem: if postmaster >> or standalone backend crashes before main(), the pop up would appear. > > When you use pg_ctl start within the command prompt, then closing the > prompt session also causes Postgres to stop immediately. Would it be a > problem? > > Another question I have is that I have seen Postgres fail to start > because of missing dependent libraries, which happened after the > postmaster startup as this was reported in the logs, could it be a > problem with this patch? >
It doesn't appear so, because we have SetErrorMode(SEM_FAILCRITICALERRORS | SEM_NOOPENFILEERRORBOX); in dlopen. -- With Regards, Amit Kapila. EnterpriseDB: http://www.enterprisedb.com