Обсуждение: Order of operations in ascii vs archive format (moderately urgent)
Is there a difference in the order of execution between an ascii dump and one using the "custom" format? Or any difference in the general operation? I need to know if I can rely on the ascii version to tell me what the custom format might have done.
Rob Sargent <robjsargent@gmail.com> writes: > Is there a difference in the order of execution between an ascii dump > and one using the "custom" format? Or any difference in the general > operation? There's not supposed to be. One standard test on the pg_dump code ispg_dump >textfilepg_dump -Fc >dumpfilepg_restore dumpfile>textfile2diff textfile textfile2 If these don't produce the same results something is broken. regards, tom lane
Wonderful news. I just ran dump and restore against same production server. Constraints and the absence of drop <object> calls appears to have saved my butt. Tom Lane wrote: > Rob Sargent <robjsargent@gmail.com> writes: > >> Is there a difference in the order of execution between an ascii dump >> and one using the "custom" format? Or any difference in the general >> operation? >> > > There's not supposed to be. One standard test on the pg_dump code is > pg_dump >textfile > pg_dump -Fc >dumpfile > pg_restore dumpfile >textfile2 > diff textfile textfile2 > If these don't produce the same results something is broken. > > regards, tom lane >
The ascii dump has serveral CREATE FUNCTION gbtreeN_{in,out} but I don't see them in the current (source) database using '\df gbtree*'. Using '\df gbt*' I get 111 functions for which all the names begin 'gbt_'. Have I lost them? The gbtreekeyN types are still there. Tom Lane wrote: > Rob Sargent <robjsargent@gmail.com> writes: > >> Is there a difference in the order of execution between an ascii dump >> and one using the "custom" format? Or any difference in the general >> operation? >> > > There's not supposed to be. One standard test on the pg_dump code is > pg_dump >textfile > pg_dump -Fc >dumpfile > pg_restore dumpfile >textfile2 > diff textfile textfile2 > If these don't produce the same results something is broken. > > regards, tom lane >
Rob Sargent <robjsargent@gmail.com> writes: > The ascii dump has serveral CREATE FUNCTION gbtreeN_{in,out} but I don't > see them in the current (source) database using '\df gbtree*'. Which ones? Pre-8.4 \df will deliberately suppress I/O functions (or functions that it thinks are I/O functions, anyway). regards, tom lane