Hello,
> > At Wed, 19 Mar 2014 19:34:10 +0900, Fujii Masao wrote
> >> > Agreed. Attached patches do that and I could "recover" the
> >> > database state with following steps,
> >>
> >> Adding new option looks like new feature rather than bug fix.
> >> I'm afraid that the backpatch of such a change to 9.3 or before
> >> is not acceptable.
> >
> > Me too. But on the other hand it simplly is a relief for the
> > consequence of the behavior of server (altough it was ill
> > operation:), and especially it is needed for at least 9.1 which
> > seems cannot be saved without it. Plus it has utterly no impact
> > on servers' behavior of any corresponding versions. So I hope it
> > is accepted.
>
> Even in 9.1, we can think that problematic situation as database corruption
> and restart the server from the backup which was successfully taken before.
> No?
Mmm. I don't think it is relevant to this problem. The problem
specific here is 'The database was running until just now, but
shutdown the master (by pacemaker), then restart, won't run
anymore'. Deleting backup_label after immediate shutdown is the
radical measure but existing system would be saved by the option.
But, honestly saying, I (also?) don't have sympathy for the
situation so much and if all or most of you think the option can
cause another problem, I won't insist about that any more.
regards,
--
Kyotaro Horiguchi
NTT Open Source Software Center