On Fri, Jan 11, 2019 at 8:01 PM Tom Lane <t...@sss.pgh.pa.us> wrote: > > Amit Kapila <amit.kapil...@gmail.com> writes: > > On Fri, Jan 11, 2019 at 5:00 AM Tom Lane <t...@sss.pgh.pa.us> wrote: > >> It's also possible that you just aren't exercising the cases where > >> trouble occurs. In particular, noting this bit in InsertOneValue(): > >> OidOutputFunctionCall(typoutput, values[i])))); > > > I have tried initdb with --debug option (If I am not wrong, it runs > > initdb under DEBUG5 mode) and didn't hit any problem after applying > > the patch. Are you expecting that we might try to open pg_proc at > > that place which can lead to the problem? > > Yes, I was concerned about regprocout in particular. It might be > okay as long as we don't try to add any regproc columns to the > BKI_BOOTSTRAP catalogs. >
Okay, I think if we get such a requirement in the future, we might want to restore the current behavior by selectively doing it for non-FSM and or non-VM forks. I will commit the patch removing these changes before committing the avoid fsm creation stuff [1] by John which is still under review. [1] - https://www.postgresql.org/message-id/CACPNZCuEijBJ9gq9jqjwvQeDB6%3Dig2egRzHhZDfTtRwhmPLVtw%40mail.gmail.com -- With Regards, Amit Kapila. EnterpriseDB: http://www.enterprisedb.com
