Re: time_t problem with postgresql

[Matteo Beccati <php%beccati.com@localhost>]

Alan Barrett ha scritto:
> On Wed, 11 Mar 2009, Daniel Carosone wrote:
>> My guess, repeated from last time[sic]: time_t is being written to
>> disk, and has changed size. This is causing fallout elsewhere, as
>> later fields are being read at the wrong offsets, giving odd results
>> like above. 
> 
> Yes, postgresql writes (used to write) time_t to disk,
> and gets confused when sizeof(time_t) changes.  See
> <https://projects.commandprompt.com/public/pgsql/changeset/27784> for
> the fix.  I don't know which versions of postgresql include this fix.

Which is only in the MAIN (8.4) branch.

Given that the OP's cluster was initialized with 8.3, the only way to
get it working would be to transfer the data files on a system with the
same characteristics to create a dump that can be restored on the newly
updated system.


Cheers

-- 
Matteo Beccati

OpenX - http://www.openx.org