On Wed, 2014-07-23 at 13:16 -0400, John Scalia wrote:
> Trying to rebuild my standby server (several times now). I know repeating something over and over while expecting a
differentresult is the definition of insanity. But, I've always
> started with:
>
> /usr/pgsql-9.3/bin/pg_basebackup -h <masters_ip_address> -D /var/lib/pgsql/9.3/data -X s -c fast
>
> I then copy my recovery.conf file into the data directory, check that everything looks correct, and try to start the
standby.The log consistently says something like:
>
> LOG: entering standby mode
> LOG: restored log file "00000006.history" from archive
> cp: cannot stat '/mnt/wallogs/archive/000000060000001200000059': No such file or directory
> LOG: restored log file "000000060000001200000059" from archive
> FATAL: requested timeline 6 is not a child of this server's history
> DETAIL: Latest checkpoint is at 12/5B000060 on timeline 5, but in the history of the requested timeline, the server
forkedoff from that timeline at 12/47000000.
> LOG: startup process (PID 60870) exited with exit code 1
> LOG: Aborting startup due to startup process failure
>
> This routine has worked for me in the past and I don't understand why it won't work now. This is supposed to be a
streamingreplication standby server. So, do I have the wrong
> flags for pg_basebackup? Or is something else going on?
Could you show your recovery.conf file? It looks like the mistake is in
there.