On Sat, Nov 24, 2012 at 3:44 PM, wd <wd@wdicc.com> wrote:
>>
>> What entries are you getting in the log file?
>>
> Logs are something like this:
>
> [ 2012-11-24 21:51:33.374 CST 583 50b0d0e5.247 1 0]LOG: database system
> was shut down in recovery at 2012-11-24 21:51:32 CST
> [ 2012-11-24 21:51:33.375 CST 583 50b0d0e5.247 2 0]LOG: starting
> point-in-time recovery to 2012-11-24 03:35:09+08
...
> [ 2012-11-24 21:51:33.591 CST 583 50b0d0e5.247 8 0]LOG: recovery
> stopping before commit of transaction 1573829, time 2012-11-24
> 03:35:17.704521+08
> [ 2012-11-24 21:51:33.591 CST 583 50b0d0e5.247 9 0]LOG: recovery has
> paused
> [ 2012-11-24 21:51:33.591 CST 583 50b0d0e5.247 10 0]HINT: Execute
> pg_xlog_replay_resume() to continue.
>
Something is strange here. There is no message about reaching a
consistent recovery point (which is why it doesn't start in hot
standby mode I guess). But when I arrange for that to happen myself,
by specifying a time before consistent recovery would be reached, I
get this error:
FATAL: requested recovery stop point is before consistent recovery point
I don't understand why are you not getting this message. I wonder if
the earlier attempts you made at doing this left the system in some
inconsistent state. If you start over again with a fresh restore of
the base backup and a recovery.conf which is correctly configured the
first time you try to use it, do you still see this?
Cheers,
Jeff