Re: pg_dump and pg_dumpall in real life
От | Craig Ringer |
---|---|
Тема | Re: pg_dump and pg_dumpall in real life |
Дата | |
Msg-id | 5281C7CA.7000407@2ndquadrant.com обсуждение исходный текст |
Ответ на | pg_dump and pg_dumpall in real life (Rafael Martinez <r.m.guerrero@usit.uio.no>) |
Список | pgsql-hackers |
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 On 11/11/2013 09:59 PM, Rafael Martinez wrote: > * We need a pg_dump solution that can generate in one step all the > necessary pieces of information needed when restoring or cloning a > database. (schema, data, privileges, users and alter database/role > data) ... and if some users/roles already exist, but have different meanings? Or some roles exist and some don't? I'm very strongly against adding and using CREATE ROLE IF NOT EXISTS. pg_restore should handle this case-by-case, forcing the user to specify explicitly role-by-role that they want a given role in the existing DB re-used if it exists, or want a new one created with a new name in case of a clash. A --rename-all-conflicting-roles and --reuse-all-conflicting-roles option could be added for the all-or-none options. IMO if neither is specified, the existence of any role name conflict should be a restore error. > * It would be great to be able to tell pg_restore that user1 in > the dump will became user2 in the restored/cloned database. The > same for the name of the database. Agreed that this would be useful. Needs to deal with the case where the users should be separated but they should remain a member of some common role, though - eg "olduser" becomes "newuser" but the dumped "olduser" was member of role "users" and "newuser" should also be member of "users", not some renamed role. - -- Craig Ringer http://www.2ndQuadrant.com/PostgreSQL Development, 24x7 Support, Training & Services -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.4.15 (GNU/Linux) Comment: Using GnuPG with Thunderbird - http://www.enigmail.net/ iQEcBAEBAgAGBQJSgcfKAAoJELBXNkqjr+S2XWIH/2c1Mcd4ldVTAPw/jAln4gNM YH8SRPlsGU0fqfbYoKg/1y0K/Wdjdlac9bjjGzYpODryaXGopf1i+pWaphF2kJTM LeMRVgEFEW7u2Dr6FXajQTQCiXLnA8C16NmmgIdqZZgYCsOwCorG+gFNfI8fZyft okCQpYcljGXzlc218DI6/o4OZBBSdLh8diTzF8+xywoXJZopdAwfHDPPpAvizPye rcUUkq1svArq78HakSuI8HoCy3ZHuiCf8mQEUPcLhFrwgh+bkrs29W7YAdD75gr4 yp32XeyOY5npXHaG9mHghs7anbUnwywJVEzpwKAf0SyPe7zunw8fdtx2NSF70no= =dusS -----END PGP SIGNATURE-----
В списке pgsql-hackers по дате отправления: