Re: trying to run PITR recovery

Warren Little <Warren(dot)Little(at)MeridiasCapital(dot)com> writes:
> I'm testing my PITR recovery procedures and something doesn't look
> right.
> The following is from the logs upon starting postgres with
> recovery.conf file in place

> @ 2007-03-23 05:57:33 MDTLOG: restored log file
> "000000010000011A000000FD" from archive
> @ 2007-03-23 05:57:35 MDTLOG: incorrect resource manager data
> checksum in record at 11A/FD492B20
> @ 2007-03-23 05:57:35 MDTLOG: redo done at 11A/FD492210

> My concern is that there were many more logfiles to be played
> following "00000010000011A000000FD"
> (ie 000000010000011E0000005C) yet it appears the recovery stop at
> that point and called it good.

Perhaps you have a bad copy of that particular log segment file?
The "incorrect checksum" failure is normal in certain circumstances
(eg, where the system crashes partway through writing out a large
WAL record), which is why the code treats it as indicating the end
of WAL. But in this case it looks more like corrupt data ...

regards, tom lane