"Jeff Lu" <jklcom@mindspring.com> writes:
> using 7.1.2
> DEBUG: redo done at (1, 3923880100)
> FATAL 2: XLogFlush: request is not satisfied
> postmaster: Startup proc 1016 exited with status 512 - abort
> What can I do to recover?
Build and run the contrib/pg_resetxlog/ utility. This will remove
the broken xlog and let you start up the database. You'll need to
look around and see if any consistency problems have been introduced
(ie, some recent transactions may have been only partially applied).
I'm guessing not, since you got as far as 'redo done', but be wary.
BTW, how did you get into this state? Do you have enough disk space
to tar up the $PGDATA directory tree for someone to look at later?
regards, tom lane