On Saturday 06 February 2010 17:32:43 Andres Freund wrote:
> On Saturday 06 February 2010 02:25:33 Tatsuo Ishii wrote:
> > Hi,
> >
> > While testing Hot Standby, I have encountered strange behavior with
> > DROP DATABASE command.
> >
> > 1) connect to "test" database at standby via psql
> > 2) issue DROP DATABASE test command to primary
> > 3) session #1 works fine
> > 4) close session #1
> > 5) "test" database dropped on standby
> >
> > Fromt the manual:
> > Running DROP DATABASE, ALTER DATABASE ... SET TABLESPACE, or ALTER
> > DATABASE ... RENAME on primary will generate a log message that will
> > cause all users connected to that database on the standby to be
> > forcibly disconnected. This action occurs immediately, whatever the
> > setting of max_standby_delay.
> >
> > So it seems at least the behavior is quite different from what the
> > docs stats. Am I missing something here?
>
> Its a small bug/typo in standby.c:ResolveRecoveryConflictWithDatabase
>
> The line:
> CancelDBBackends(dbid, PROCSIG_RECOVERY_CONFLICT_TABLESPACE, true);
For the case it should not be clear, the reason that
PROCSIG_RECOVERY_CONFLICT_TABLESPACE did not kill the session is that
currently all tablespace conflicts are valid only inside a transaction, so when
receiving the recovery conflict checks whether its not inside a transaction
block "anymore" and continues happily.
Andres