Re: Standalone backends run StartupXLOG in an incorrect environment

On Monday 21 September 2009 14:24:07 Tom Lane wrote:
> Alvaro Herrera <alvherre(at)commandprompt(dot)com> writes:
> > So if you need to enter standalone mode, you'll have to start
> > postmaster, wait for replay to finish, stop it and restart standalone.
>
> Yeah, that's the case at the moment.
>
> > Would this be a problem when you need standalone mode in an emergency,
> > for example when the database won't start due to Xid wraparound?
>
> If it won't run through StartupXLOG under the postmaster, it's not going
> to do so standalone either.
>

Hmm... istr cases where I couldn't startup regular postgres but could in
stand-alone mode that had system indexes disabled...I could be misremembering
that so that the postmaster would start, I just couldn't connect unless in
stand-alone. In any case this does seem less than ideal, but if there aren't
any better options...

--
Robert Treat
Conjecture: http://www.xzilla.net
Consulting: http://www.omniti.com