Обсуждение: HELP pls
Hello, help me please, I don't know how, but i got 2 postgres users. How can I delete the first one ???? thx alot for your answers. Oleg
spol@mail.ru wrote: > Hello, help me please, > I don't know how, but i got 2 postgres users. > How can I delete the first one ???? > thx alot for your answers. You've got transaction wraparound problems. Start here: http://search.postgresql.org/search?q=transaction+wraparound&m=1&l=&d=365&s=r -- Alvaro Herrera http://www.CommandPrompt.com/ The PostgreSQL Company - Command Prompt, Inc.
Alvaro Herrera <alvherre@commandprompt.com> writes: > spol@mail.ru wrote: >> I don't know how, but i got 2 postgres users. > You've got transaction wraparound problems. Start here: Vacuuming isn't gonna help though. The tuple at (0,2) has been frozen, which means that vacuum has certainly been run on the table since the update; so he's wrapped around so far that vacuum doesn't think it can remove the (0,1) tuple. I assume the OP is just trying to dump this database so he can upgrade to something remotely modern. It might work to change the second tuple's usesysid to something else (eg, 2) so that pg_dump doesn't see multiple rows matching other objects' owner fields. regards, tom lane
how to dump it ??? pg_dump is not working : pg_dump: query to obtain list of schemas failed: ERROR: more than one row returned by a subquery used as an expression Tom Lane ïèøåò: > Alvaro Herrera <alvherre@commandprompt.com> writes: > >> spol@mail.ru wrote: >> >>> I don't know how, but i got 2 postgres users. >>> > > >> You've got transaction wraparound problems. Start here: >> > > Vacuuming isn't gonna help though. The tuple at (0,2) has been frozen, > which means that vacuum has certainly been run on the table since the > update; so he's wrapped around so far that vacuum doesn't think it can > remove the (0,1) tuple. > > I assume the OP is just trying to dump this database so he can upgrade > to something remotely modern. It might work to change the second > tuple's usesysid to something else (eg, 2) so that pg_dump doesn't > see multiple rows matching other objects' owner fields. > > regards, tom lane > > >